Hyunjung Shim

Seojeong Park, Jiho Choi, Junyong Kang et al.

Recent multimodal large language models have demonstrated strong reasoning ability, yet their reliability as automated evaluators remains limited by a critical weakness: when visual evidence conflicts with textual cues, MLLM judges tend to reward plausible narratives over perceptually correct answers. We identify and systematically analyze this phenomenon, which we term Perceptual Judgment Bias. Through controlled visual perturbations, existing multimodal judges frequently anchor on the response text instead of their own visual perception, leading to inconsistent and non-verifiable evaluations. To address this issue, we introduce the Perceptually Perturbed Judgment Dataset, which constructs minimally edited counterfactual responses that isolate perceptual errors and enable verifiable supervision. Building on this dataset, we develop a unified training framework that combines a structured GRPO-based reward with a batch-ranking objective, achieving coherent global ordering without explicit pairwise labels. Experiments across diverse MLLM-as-a-Judge benchmarks show that our approach substantially improves perceptual fidelity, ranking coherence, and alignment with human evaluation. Our results establish a scalable and generalizable pathway for training multimodal judges that are perceptually grounded, interpretable, and robust to visual-reasoning conflicts.

Junsung Park, Kyungmin Kim, Hyunjung Shim

Existing LiDAR semantic segmentation methods often struggle with performance declines in adverse weather conditions. Previous work has addressed this issue by simulating adverse weather or employing universal data augmentation during training. However, these methods lack a detailed analysis and understanding of how adverse weather negatively affects LiDAR semantic segmentation performance. Motivated by this issue, we identified key factors of adverse weather and conducted a toy experiment to pinpoint the main causes of performance degradation: (1) Geometric perturbation due to refraction caused by fog or droplets in the air and (2) Point drop due to energy absorption and occlusions. Based on these findings, we propose new strategic data augmentation techniques. First, we introduced a Selective Jittering (SJ) that jitters points in the random range of depth (or angle) to mimic geometric perturbation. Additionally, we developed a Learnable Point Drop (LPD) to learn vulnerable erase patterns with a Deep Q-Learning Network to approximate the point drop phenomenon from adverse weather conditions. Without precise weather simulation, these techniques strengthen the LiDAR semantic segmentation model by exposing it to vulnerable conditions identified by our data-centric analysis. Experimental results confirmed the suitability of the proposed data augmentation methods for enhancing robustness against adverse weather conditions. Our method achieves a notable 39.5 mIoU on the SemanticKITTI-to-SemanticSTF benchmark, improving the baseline by 8.1\%p and establishing a new state-of-the-art. Our code will be released at \url{https://github.com/engineerJPark/LiDARWeather}.

Kyungjune Baek, Hyunjung Shim

Transfer learning for GANs successfully improves generation performance under low-shot regimes. However, existing studies show that the pretrained model using a single benchmark dataset is not generalized to various target datasets. More importantly, the pretrained model can be vulnerable to copyright or privacy risks as membership inference attack advances. To resolve both issues, we propose an effective and unbiased data synthesizer, namely Primitives-PS, inspired by the generic characteristics of natural images. Specifically, we utilize 1) the generic statistics on the frequency magnitude spectrum, 2) the elementary shape (i.e., image composition via elementary shapes) for representing the structure information, and 3) the existence of saliency as prior. Since our synthesizer only considers the generic properties of natural images, the single model pretrained on our dataset can be consistently transferred to various target datasets, and even outperforms the previous methods pretrained with the natural images in terms of Fr'echet inception distance. Extensive analysis, ablation study, and evaluations demonstrate that each component of our data synthesizer is effective, and provide insights on the desirable nature of the pretrained model for the transferability of GANs.

Seonho Lee, Jiho Choi, Seohyun Lim et al.

Recent advancements in text-to-image diffusion models have demonstrated remarkable success, yet they often struggle to fully capture the user's intent. Existing approaches using textual inputs combined with bounding boxes or region masks fall short in providing precise spatial guidance, often leading to misaligned or unintended object orientation. To address these limitations, we propose Scribble-Guided Diffusion (ScribbleDiff), a training-free approach that utilizes simple user-provided scribbles as visual prompts to guide image generation. However, incorporating scribbles into diffusion models presents challenges due to their sparse and thin nature, making it difficult to ensure accurate orientation alignment. To overcome these challenges, we introduce moment alignment and scribble propagation, which allow for more effective and flexible alignment between generated images and scribble inputs. Experimental results on the PASCAL-Scribble dataset demonstrate significant improvements in spatial control and consistency, showcasing the effectiveness of scribble-based guidance in diffusion models. Our code is available at https://github.com/kaist-cvml-lab/scribble-diffusion.

Jaeyo Shin, Jiwook Kim, Hyunjung Shim

Representation Alignment (REPA) has emerged as a simple way to accelerate Diffusion Transformers training in latent space. At the same time, pixel-space diffusion transformers such as Just image Transformers (JiT) have attracted growing attention because they remove a dependency on a pretrained tokenizer, and then avoid the reconstruction bottleneck of latent diffusion. This paper shows that the REPA can fail for JiT. REPA yields worse FID for JiT as training proceeds and collapses diversity on image subsets that are tightly clustered in the representation space of pretrained semantic encoder on ImageNet. We trace the failure to an information asymmetry: denoising occurs in the high dimensional image space, while the semantic target is strongly compressed, making direct regression a shortcut objective. We propose PixelREPA, which transforms the alignment target and constrains alignment with a Masked Transformer Adapter that combines a shallow transformer adapter with partial token masking. PixelREPA improves both training convergence and final quality. PixelREPA reduces FID from 3.66 to 3.17 for JiT-B$/16$ and improves Inception Score (IS) from 275.1 to 284.6 on ImageNet $256 \times 256$, while achieving $> 2\times$ faster convergence. Finally, PixelREPA-H$/16$ achieves FID$=1.81$ and IS$=317.2$. Our code is available at https://github.com/kaist-cvml/PixelREPA.

Junsung Park, Hyunjung Shim

Existing Driving VLAs predict trajectories while largely ignoring their visual tokens -- a phenomenon we trace not to insufficient training but to a structurally ill-posed task formulation. We show that trajectory recovery, when viewed through the lens of inverse kinematics, requires both a current and a future visual state as boundary conditions; existing VLAs supply only the former, which encourages the model to shortcut through ego status and text commands alone. To address this, we re-design Driving VLA in the style of an inverse kinematics solver. First, a next visual state prediction objective that requires the LLM to predict the future visual scene provides dense visual supervision and suppresses shortcut paths. Second, a separate Inverse Kinematics Network (a cross-attention-based conditional diffusion model) that takes only the current and future visual states as input is designed to suppress reliance on ego status and textual shortcuts during trajectory decoding. With this simple prescription alone, our 0.5B-scale model recovers visual grounding and reaches trajectory planning performance comparable to 7B--8B VLAs more than an order of magnitude larger, on both the closed-loop NAVSIM-v2 and the nuScenes benchmarks. Extensive analysis further shows that this improvement stems from a recovered ability to exploit visual features, with the effect being most pronounced in dynamic driving situations such as turning.

Dongseob Kim, Seungho Lee, Junsuk Choe et al.

State-of-the-art techniques in weakly-supervised semantic segmentation (WSSS) using image-level labels exhibit severe performance degradation on driving scene datasets such as Cityscapes. To address this challenge, we develop a new WSSS framework tailored to driving scene datasets. Based on extensive analysis of dataset characteristics, we employ Contrastive Language-Image Pre-training (CLIP) as our baseline to obtain pseudo-masks. However, CLIP introduces two key challenges: (1) pseudo-masks from CLIP lack in representing small object classes, and (2) these masks contain notable noise. We propose solutions for each issue as follows. (1) We devise Global-Local View Training that seamlessly incorporates small-scale patches during model training, thereby enhancing the model's capability to handle small-sized yet critical objects in driving scenes (e.g., traffic light). (2) We introduce Consistency-Aware Region Balancing (CARB), a novel technique that discerns reliable and noisy regions through evaluating the consistency between CLIP masks and segmentation predictions. It prioritizes reliable pixels over noisy pixels via adaptive loss weighting. Notably, the proposed method achieves 51.8\% mIoU on the Cityscapes test dataset, showcasing its potential as a strong WSSS baseline on driving scene datasets. Experimental results on CamVid and WildDash2 demonstrate the effectiveness of our method across diverse datasets, even with small-scale datasets or visually challenging conditions. The code is available at https://github.com/k0u-id/CARB.

Soyeon Yoon, Chang Wook Seo, Hyunjung Shim

Learning dense correspondences across deformable 3D shapes remains a long-standing challenge due to structural variability, non-isometric deformation, and inconsistent topology. Existing methods typically trade off generalization, geometric fidelity, and efficiency. We address this by proposing SGSoft, a unified intrinsic pipeline that (i) constructs a geodesic correspondence field on a canonical template, (ii) learns multimodal dense descriptors guided by pretrained semantic priors with this geodesic correspondence field supervision, (iii) retrieves dense correspondences in a single feed-forward pass via nearest-neighbor search in descriptor space. This formulation enables stable and topology-invariant supervision under large pose variation, structural differences, and remeshing. SGSoft achieves state-of-the-art inter-category generalization while offering the best accuracy-efficiency trade-off among prior methods. It also achieves near real-time inference without pre-alignment, pairwise optimization, or post-refinement. Learned descriptors can be transferred effectively to downstream tasks such as semantic segmentation and deformation transfer, establishing a scalable and deployment-ready paradigm for dense 3D correspondence.

NaHyeon Park, Kunhee Kim, Junsuk Choe et al.

Rich feature representations derived from CLIP-ViT have been widely utilized in AI-generated image detection. While most existing methods primarily leverage features from the final layer, we systematically analyze the contributions of layer-wise features to this task. Our study reveals that earlier layers provide more localized and generalizable features, often surpassing the performance of final-layer features in detection tasks. Moreover, we find that different layers capture distinct aspects of the data, each contributing uniquely to AI-generated image detection. Motivated by these findings, we introduce a novel adaptive method, termed MoLD, which dynamically integrates features from multiple ViT layers using a gating-based mechanism. Extensive experiments on both GAN- and diffusion-generated images demonstrate that MoLD significantly improves detection performance, enhances generalization across diverse generative models, and exhibits robustness in real-world scenarios. Finally, we illustrate the scalability and versatility of our approach by successfully applying it to other pre-trained ViTs, such as DINOv2.

NaHyeon Park, Kunhee Kim, Hyunjung Shim

Recent breakthroughs in text-to-image models have opened up promising research avenues in personalized image generation, enabling users to create diverse images of a specific subject using natural language prompts. However, existing methods often suffer from performance degradation when given only a single reference image. They tend to overfit the input, producing highly similar outputs regardless of the text prompt. This paper addresses the challenge of one-shot personalization by mitigating overfitting, enabling the creation of controllable images through text prompts. Specifically, we propose a selective fine-tuning strategy that focuses on the text encoder. Furthermore, we introduce three key techniques to enhance personalization performance: (1) augmentation tokens to encourage feature disentanglement and alleviate overfitting, (2) a knowledge-preservation loss to reduce language drift and promote generalizability across diverse prompts, and (3) SNR-weighted sampling for efficient training. Extensive experiments demonstrate that our approach efficiently generates high-quality, diverse images using only a single reference image while significantly reducing memory and storage requirements.

Jiwook Kim, Seonho Lee, Jaeyo Shin et al.

Score distillation sampling (SDS) has emerged as an effective framework in text-driven 3D editing tasks, leveraging diffusion models for 3D-consistent editing. However, existing SDS-based 3D editing methods suffer from long training times and produce low-quality results. We identify that the root cause of this performance degradation is \textit{their conflict with the sampling dynamics of diffusion models}. Addressing this conflict allows us to treat SDS as a diffusion reverse process for 3D editing via sampling from data space. In contrast, existing methods naively distill the score function using diffusion models. From these insights, we propose DreamCatalyst, a novel framework that considers these sampling dynamics in the SDS framework. Specifically, we devise the optimization process of our DreamCatalyst to approximate the diffusion reverse process in editing tasks, thereby aligning with diffusion sampling dynamics. As a result, DreamCatalyst successfully reduces training time and improves editing quality. Our method offers two modes: (1) a fast mode that edits Neural Radiance Fields (NeRF) scenes approximately 23 times faster than current state-of-the-art NeRF editing methods, and (2) a high-quality mode that produces superior results about 8 times faster than these methods. Notably, our high-quality mode outperforms current state-of-the-art NeRF editing methods in terms of both speed and quality. DreamCatalyst also surpasses the state-of-the-art 3D Gaussian Splatting (3DGS) editing methods, establishing itself as an effective and model-agnostic 3D editing solution. See more extensive results on our project page: https://dream-catalyst.github.io.

Hyogon Ryu, NaHyeon Park, Hyunjung Shim

Despite the widespread use of text-to-image diffusion models across various tasks, their computational and memory demands limit practical applications. To mitigate this issue, quantization of diffusion models has been explored. It reduces memory usage and computational costs by compressing weights and activations into lower-bit formats. However, existing methods often struggle to preserve both image quality and text-image alignment, particularly in lower-bit($<$ 8bits) quantization. In this paper, we analyze the challenges associated with quantizing text-to-image diffusion models from a distributional perspective. Our analysis reveals that activation outliers play a crucial role in determining image quality. Additionally, we identify distinctive patterns in cross-attention scores, which significantly affects text-image alignment. To address these challenges, we propose Distribution-aware Group Quantization (DGQ), a method that identifies and adaptively handles pixel-wise and channel-wise outliers to preserve image quality. Furthermore, DGQ applies prompt-specific logarithmic quantization scales to maintain text-image alignment. Our method demonstrates remarkable performance on datasets such as MS-COCO and PartiPrompts. We are the first to successfully achieve low-bit quantization of text-to-image diffusion models without requiring additional fine-tuning of weight quantization parameters. Code is available at https://github.com/ugonfor/DGQ.

Youngsun Lim, Hojun Choi, Hyunjung Shim

Despite the impressive success of text-to-image (TTI) generation models, existing studies overlook the issue of whether these models accurately convey factual information. In this paper, we focus on the problem of image hallucination, where images created by generation models fail to faithfully depict factual content. To address this, we introduce I-HallA (Image Hallucination evaluation with Question Answering), a novel automated evaluation metric that measures the factuality of generated images through visual question answering (VQA). We also introduce I-HallA v1.0, a curated benchmark dataset for this purpose. As part of this process, we develop a pipeline that generates high-quality question-answer pairs using multiple GPT-4 Omni-based agents, with human judgments to ensure accuracy. Our evaluation protocols measure image hallucination by testing if images from existing TTI models can correctly respond to these questions. The I-HallA v1.0 dataset comprises 1.2K diverse image-text pairs across nine categories with 1,000 rigorously curated questions covering various compositional challenges. We evaluate five TTI models using I-HallA and reveal that these state-of-the-art models often fail to accurately convey factual information. Moreover, we validate the reliability of our metric by demonstrating a strong Spearman correlation ($ρ$=0.95) with human judgments. We believe our benchmark dataset and metric can serve as a foundation for developing factually accurate TTI generation models. Additional resources can be found on our project page: https://sgt-lim.github.io/I-HallA/.

Youngsun Lim, Hyunjung Shim

Text-to-image generation has shown remarkable progress with the emergence of diffusion models. However, these models often generate factually inconsistent images, failing to accurately reflect the factual information and common sense conveyed by the input text prompts. We refer to this issue as Image hallucination. Drawing from studies on hallucinations in language models, we classify this problem into three types and propose a methodology that uses factual images retrieved from external sources to generate realistic images. Depending on the nature of the hallucination, we employ off-the-shelf image editing tools, either InstructPix2Pix or IP-Adapter, to leverage factual information from the retrieved image. This approach enables the generation of images that accurately reflect the facts and common sense.

Minhyun Lee, Song Park, Byeongho Heo et al.

Recent advancements in Deep Neural Network (DNN) models have significantly improved performance across computer vision tasks. However, achieving highly generalizable and high-performing vision models requires expansive datasets, resulting in significant storage requirements. This storage challenge is a critical bottleneck for scaling up models. A recent breakthrough by SeiT proposed the use of Vector-Quantized (VQ) feature vectors (i.e., tokens) as network inputs for vision classification. This approach achieved 90% of the performance of a model trained on full-pixel images with only 1% of the storage. While SeiT needs labeled data, its potential in scenarios beyond fully supervised learning remains largely untapped. In this paper, we extend SeiT by integrating Masked Token Modeling (MTM) for self-supervised pre-training. Recognizing that self-supervised approaches often demand more data due to the lack of labels, we introduce TokenAdapt and ColorAdapt. These methods facilitate comprehensive token-friendly data augmentation, effectively addressing the increased data requirements of self-supervised learning. We evaluate our approach across various scenarios, including storage-efficient ImageNet-1k classification, fine-grained classification, ADE-20k semantic segmentation, and robustness benchmarks. Experimental results demonstrate consistent performance improvement in diverse experiments, validating the effectiveness of our method. Code is available at https://github.com/naver-ai/seit.

NaHyeon Park, Namin An, Kunhee Kim et al.

Large vision-language model (LVLM) based text-to-image (T2I) systems have become the dominant paradigm in image generation, yet whether they amplify social biases remains insufficiently understood. In this paper, we show that LVLM-based models produce markedly more socially biased images than non-LVLM-based models. We introduce a 1,024 prompt benchmark spanning four levels of linguistic complexity and evaluate demographic bias across multiple attributes in a systematic manner. Our analysis identifies system prompts, the predefined instructions guiding LVLMs, as a primary driver of biased behavior. Through decoded intermediate representations, token-probability diagnostics, and embedding-association analyses, we reveal how system prompts encode demographic priors that propagate into image synthesis. To this end, we propose FairPro, a training-free meta-prompting framework that enables LVLMs to self-audit and construct fairness-aware system prompts at test time. Experiments on two LVLM-based T2I models, SANA and Qwen-Image, show that FairPro substantially reduces demographic bias while preserving text-image alignment. We believe our findings provide deeper insight into the central role of system prompts in bias propagation and offer a practical, deployable approach for building more socially responsible T2I systems.

Seojeong Park, Jiho Choi, Kyungjune Baek et al.

Video Moment Retrieval (MR) aims to localize moments within a video based on a given natural language query. Given the prevalent use of platforms like YouTube for information retrieval, the demand for MR techniques is significantly growing. Recent DETR-based models have made notable advances in performance but still struggle with accurately localizing short moments. Through data analysis, we identified limited feature diversity in short moments, which motivated the development of MomentMix. MomentMix employs two augmentation strategies: ForegroundMix and BackgroundMix, each enhancing the feature representations of the foreground and background, respectively. Additionally, our analysis of prediction bias revealed that short moments particularly struggle with accurately predicting their center positions of moments. To address this, we propose a Length-Aware Decoder, which conditions length through a novel bipartite matching process. Our extensive studies demonstrate the efficacy of our length-aware approach, especially in localizing short moments, leading to improved overall performance. Our method surpasses state-of-the-art DETR-based methods on benchmark datasets, achieving the highest R1 and mAP on QVHighlights and the highest R1@0.7 on TACoS and Charades-STA (such as a 2.46% gain in R1@0.7 and a 2.57% gain in mAP average for QVHighlights). The code is available at https://github.com/sjpark5800/LA-DETR.

Minhyun Lee, Seungho Lee, Song Park et al.

Referring Image Segmentation (RIS) is an advanced vision-language task that involves identifying and segmenting objects within an image as described by free-form text descriptions. While previous studies focused on aligning visual and language features, exploring training techniques, such as data augmentation, remains underexplored. In this work, we explore effective data augmentation for RIS and propose a novel training framework called Masked Referring Image Segmentation (MaskRIS). We observe that the conventional image augmentations fall short of RIS, leading to performance degradation, while simple random masking significantly enhances the performance of RIS. MaskRIS uses both image and text masking, followed by Distortion-aware Contextual Learning (DCL) to fully exploit the benefits of the masking strategy. This approach can improve the model's robustness to occlusions, incomplete information, and various linguistic complexities, resulting in a significant performance improvement. Experiments demonstrate that MaskRIS can easily be applied to various RIS models, outperforming existing methods in both fully supervised and weakly supervised settings. Finally, MaskRIS achieves new state-of-the-art performance on RefCOCO, RefCOCO+, and RefCOCOg datasets. Code is available at https://github.com/naver-ai/maskris.

Seoungyoon Kang, Youngsun Lim, Hyunjung Shim

Traditional dataset distillation primarily focuses on image representation while often overlooking the important role of labels. In this study, we introduce Label-Augmented Dataset Distillation (LADD), a new dataset distillation framework enhancing dataset distillation with label augmentations. LADD sub-samples each synthetic image, generating additional dense labels to capture rich semantics. These dense labels require only a 2.5% increase in storage (ImageNet subsets) with significant performance benefits, providing strong learning signals. Our label generation strategy can complement existing dataset distillation methods for significantly enhancing their training efficiency and performance. Experimental results demonstrate that LADD outperforms existing methods in terms of computational overhead and accuracy. With three high-performance dataset distillation algorithms, LADD achieves remarkable gains by an average of 14.9% in accuracy. Furthermore, the effectiveness of our method is proven across various datasets, distillation hyperparameters, and algorithms. Finally, our method improves the cross-architecture robustness of the distilled dataset, which is important in the application scenario.

Seungjun Yu, Seonho Lee, Namho Kim et al.

Recent advancements in multimodal large language models (MLLMs) have shown strong understanding of driving scenes, drawing interest in their application to autonomous driving. However, high-level reasoning in safety-critical scenarios, where avoiding one traffic risk can create another, remains a major challenge. Such reasoning is often infeasible with only a single front view and requires a comprehensive view of the environment, which we achieve through multi-view inputs. We define Safety-Critical Reasoning as a new task that leverages multi-view inputs to address this challenge. Then, we distill Safety-Critical Reasoning into two stages: first resolve the immediate risk, then mitigate the decision-induced downstream risks. To support this, we introduce WaymoQA, a dataset of 35,000 human-annotated question-answer pairs covering complex, high-risk driving scenarios. The dataset includes multiple-choice and open-ended formats across both image and video modalities. Experiments reveal that existing MLLMs underperform in safety-critical scenarios compared to normal scenes, but fine-tuning with WaymoQA significantly improves their reasoning ability, highlighting the effectiveness of our dataset in developing safer and more reasoning-capable driving agents. Our code and data are provided in https://github.com/sjyu001/WaymoQA

Dongseob Kim, Hyunjung Shim

Multi-label classification is crucial for comprehensive image understanding, yet acquiring accurate annotations is challenging and costly. To address this, a recent study suggests exploiting unsupervised multi-label classification leveraging CLIP, a powerful vision-language model. Despite CLIP's proficiency, it suffers from view-dependent predictions and inherent bias, limiting its effectiveness. We propose a novel method that addresses these issues by leveraging multiple views near target objects, guided by Class Activation Mapping (CAM) of the classifier, and debiasing pseudo-labels derived from CLIP predictions. Our Classifier-guided CLIP Distillation (CCD) enables selecting multiple local views without extra labels and debiasing predictions to enhance classification performance. Experimental results validate our method's superiority over existing techniques across diverse datasets. The code is available at https://github.com/k0u-id/CCD.

Song Park, Sanghyuk Chun, Junbum Cha et al.

Automatic few-shot font generation aims to solve a well-defined, real-world problem because manual font designs are expensive and sensitive to the expertise of designers. Existing methods learn to disentangle style and content elements by developing a universal style representation for each font style. However, this approach limits the model in representing diverse local styles, because it is unsuitable for complicated letter systems, for example, Chinese, whose characters consist of a varying number of components (often called "radical") -- with a highly complex structure. In this paper, we propose a novel font generation method that learns localized styles, namely component-wise style representations, instead of universal styles. The proposed style representations enable the synthesis of complex local details in text designs. However, learning component-wise styles solely from a few reference glyphs is infeasible when a target script has a large number of components, for example, over 200 for Chinese. To reduce the number of required reference glyphs, we represent component-wise styles by a product of component and style factors, inspired by low-rank matrix factorization. Owing to the combination of strong representation and a compact factorization strategy, our method shows remarkably better few-shot font generation results (with only eight reference glyphs) than other state-of-the-art methods. Moreover, strong locality supervision, for example, location of each component, skeleton, or strokes, was not utilized. The source code is available at https://github.com/clovaai/lffont and https://github.com/clovaai/fewshot-font-generation.

Song Park, Sanghyuk Chun, Junbum Cha et al.

A few-shot font generation (FFG) method has to satisfy two objectives: the generated images should preserve the underlying global structure of the target character and present the diverse local reference style. Existing FFG methods aim to disentangle content and style either by extracting a universal representation style or extracting multiple component-wise style representations. However, previous methods either fail to capture diverse local styles or cannot be generalized to a character with unseen components, e.g., unseen language systems. To mitigate the issues, we propose a novel FFG method, named Multiple Localized Experts Few-shot Font Generation Network (MX-Font). MX-Font extracts multiple style features not explicitly conditioned on component labels, but automatically by multiple experts to represent different local concepts, e.g., left-side sub-glyph. Owing to the multiple experts, MX-Font can capture diverse local concepts and show the generalizability to unseen languages. During training, we utilize component labels as weak supervision to guide each expert to be specialized for different local concepts. We formulate the component assign problem to each expert as the graph matching problem, and solve it by the Hungarian algorithm. We also employ the independence loss and the content-style adversarial loss to impose the content-style disentanglement. In our experiments, MX-Font outperforms previous state-of-the-art FFG methods in the Chinese generation and cross-lingual, e.g., Chinese to Korean, generation. Source code is available at https://github.com/clovaai/mxfont.

Song Park, Sanghyuk Chun, Junbum Cha et al.

Automatic few-shot font generation is a practical and widely studied problem because manual designs are expensive and sensitive to the expertise of designers. Existing few-shot font generation methods aim to learn to disentangle the style and content element from a few reference glyphs, and mainly focus on a universal style representation for each font style. However, such approach limits the model in representing diverse local styles, and thus makes it unsuitable to the most complicated letter system, e.g., Chinese, whose characters consist of a varying number of components (often called "radical") with a highly complex structure. In this paper, we propose a novel font generation method by learning localized styles, namely component-wise style representations, instead of universal styles. The proposed style representations enable us to synthesize complex local details in text designs. However, learning component-wise styles solely from reference glyphs is infeasible in the few-shot font generation scenario, when a target script has a large number of components, e.g., over 200 for Chinese. To reduce the number of reference glyphs, we simplify component-wise styles by a product of component factor and style factor, inspired by low-rank matrix factorization. Thanks to the combination of strong representation and a compact factorization strategy, our method shows remarkably better few-shot font generation results (with only 8 reference glyph images) than other state-of-the-arts, without utilizing strong locality supervision, e.g., location of each component, skeleton, or strokes. The source code is available at https://github.com/clovaai/lffont.

Kunhee Kim, NaHyeon Park, Kibeom Hong et al.

Textual Inversion (TI) is an efficient approach to text-to-image personalization but often fails on complex prompts. We trace these failures to embedding norm inflation: learned tokens drift to out-of-distribution magnitudes, degrading prompt conditioning in pre-norm Transformers. Empirically, we show semantics are primarily encoded by direction in CLIP token space, while inflated norms harm contextualization; theoretically, we analyze how large magnitudes attenuate positional information and hinder residual updates in pre-norm blocks. We propose Directional Textual Inversion (DTI), which fixes the embedding magnitude to an in-distribution scale and optimizes only direction on the unit hypersphere via Riemannian SGD. We cast direction learning as MAP with a von Mises-Fisher prior, yielding a constant-direction prior gradient that is simple and efficient to incorporate. Across personalization tasks, DTI improves text fidelity over TI and TI-variants while maintaining subject similarity. Crucially, DTI's hyperspherical parameterization enables smooth, semantically coherent interpolation between learned concepts (slerp), a capability that is absent in standard TI. Our findings suggest that direction-only optimization is a robust and scalable path for prompt-faithful personalization.

Hyogon Ryu, Seohyun Lim, Hyunjung Shim

The emergence of billion-parameter diffusion models such as Stable Diffusion XL, Imagen, and DALL-E 3 has significantly propelled the domain of generative AI. However, their large-scale architecture presents challenges in fine-tuning and deployment due to high resource demands and slow inference speed. This paper explores the relatively unexplored yet promising realm of fine-tuning quantized diffusion models. Our analysis revealed that the baseline neglects the distinct patterns in model weights and the different roles throughout time steps when finetuning the diffusion model. To address these limitations, we introduce a novel memory-efficient fine-tuning method specifically designed for quantized diffusion models, dubbed TuneQDM. Our approach introduces quantization scales as separable functions to consider inter-channel weight patterns. Then, it optimizes these scales in a timestep-specific manner for effective reflection of the role of each time step. TuneQDM achieves performance on par with its full-precision counterpart while simultaneously offering significant memory efficiency. Experimental results demonstrate that our method consistently outperforms the baseline in both single-/multi-subject generations, exhibiting high subject fidelity and prompt fidelity comparable to the full precision model.

Seonho Lee, Jiho Choi, Inha Kang et al.

Vision-Language Models (VLMs) have shown remarkable performance on diverse visual and linguistic tasks, yet they remain fundamentally limited in their understanding of 3D spatial structures. We propose Geometric Distillation, a lightweight, annotation-free fine-tuning framework that injects human-inspired geometric cues into pretrained VLMs without modifying their architecture. By distilling (1) sparse correspondences, (2) relative depth relations, and (3) dense cost volumes from off-the-shelf 3D foundation models (e.g., MASt3R, VGGT), our method shapes representations to be geometry-aware while remaining compatible with natural image-text inputs. Through extensive evaluations on 3D vision-language reasoning and 3D perception benchmarks, our method consistently outperforms prior approaches, achieving improved 3D spatial reasoning with significantly lower computational cost. Our work demonstrates a scalable and efficient path to bridge 2D-trained VLMs with 3D understanding, opening up wider use in spatially grounded multimodal tasks.

Junsung Park, Hwijeong Lee, Inha Kang et al.

Existing domain generalization methods for LiDAR semantic segmentation under adverse weather struggle to accurately predict "things" categories compared to "stuff" categories. In typical driving scenes, "things" categories can be dynamic and associated with higher collision risks, making them crucial for safe navigation and planning. Recognizing the importance of "things" categories, we identify their performance drop as a serious bottleneck in existing approaches. We observed that adverse weather induces degradation of semantic-level features and both corruption of local features, leading to a misprediction of "things" as "stuff". To mitigate these corruptions, we suggest our method, NTN - segmeNt Things for No-accident. To address semantic-level feature corruption, we bind each point feature to its superclass, preventing the misprediction of things classes into visually dissimilar categories. Additionally, to enhance robustness against local corruption caused by adverse weather, we define each LiDAR beam as a local region and propose a regularization term that aligns the clean data with its corrupted counterpart in feature space. NTN achieves state-of-the-art performance with a +2.6 mIoU gain on the SemanticKITTI-to-SemanticSTF benchmark and +7.9 mIoU on the SemanticPOSS-to-SemanticSTF benchmark. Notably, NTN achieves a +4.8 and +7.9 mIoU improvement on "things" classes, respectively, highlighting its effectiveness.

Jiho Choi, Seonho Lee, Minhyun Lee et al.

Open-Vocabulary Part Segmentation (OVPS) is an emerging field for recognizing fine-grained parts in unseen categories. We identify two primary challenges in OVPS: (1) the difficulty in aligning part-level image-text correspondence, and (2) the lack of structural understanding in segmenting object parts. To address these issues, we propose PartCATSeg, a novel framework that integrates object-aware part-level cost aggregation, compositional loss, and structural guidance from DINO. Our approach employs a disentangled cost aggregation strategy that handles object and part-level costs separately, enhancing the precision of part-level segmentation. We also introduce a compositional loss to better capture part-object relationships, compensating for the limited part annotations. Additionally, structural guidance from DINO features improves boundary delineation and inter-part understanding. Extensive experiments on Pascal-Part-116, ADE20K-Part-234, and PartImageNet datasets demonstrate that our method significantly outperforms state-of-the-art approaches, setting a new baseline for robust generalization to unseen part categories.

Na Min An, Eunki Kim, James Thorne et al.

Contrastive Language-Image Pretraining (CLIP) enables zero-shot inference in downstream tasks such as image-text retrieval and classification. However, recent works extending CLIP suffer from the issue of modality gap, which arises when the image and text embeddings are projected to disparate manifolds, deviating from the intended objective of image-text contrastive learning. We discover that this phenomenon is linked to the modality-specific characteristic that each image/text encoder independently possesses and propose two methods to address the modality gap: (1) a post-hoc embedding standardization method, $\text{I0T}_{\text{post}}$ that reduces the modality gap approximately to zero and (2) a trainable method, $\text{I0T}_{\text{async}}$, to alleviate the modality gap problem by adding two normalization layers for each encoder. Our I0T framework can significantly reduce the modality gap while preserving the original embedding representations of trained models with their locked parameters. In practice, $\text{I0T}_{\text{post}}$ can serve as an alternative explainable automatic evaluation metric of widely used CLIPScore (CLIP-S).

Na Min An, Eunki Kim, Wan Ju Kang et al.

For individuals with blindness or low vision (BLV), navigating complex environments can pose serious risks. Large Vision-Language Models (LVLMs) show promise for generating scene descriptions, but their effectiveness for BLV users remains underexplored. To address this gap, we conducted a user study with eight BLV participants to systematically evaluate preferences for six types of LVLM descriptions. While they helped to reduce fear and improve actionability, user ratings showed wide variation in sufficiency and conciseness. Furthermore, GPT-4o--despite its strong potential to refine descriptions--was not consistently preferred by participants. We use the insights obtained from the user study to build training data for building our new automatic evaluation metric that can capture BLV preferences effectively. Our findings underscore the urgent need for BLV-centered evaluation metrics and human-in-the-loop feedback to advance LVLM description quality for accessibility.

Yunji Jung, Seokju Lee, Tair Djanibekov et al.

Text-guided non-rigid editing involves complex edits for input images, such as changing motion or compositions within their surroundings. Since it requires manipulating the input structure, existing methods often struggle with preserving object identity and background, particularly when combined with Stable Diffusion. In this work, we propose a training-free approach for non-rigid editing with Stable Diffusion, aimed at improving the identity preservation quality without compromising editability. Our approach comprises three stages: text optimization, latent inversion, and timestep-aware text injection sampling. Inspired by the success of Imagic, we employ their text optimization for smooth editing. Then, we introduce latent inversion to preserve the input image's identity without additional model fine-tuning. To fully utilize the input reconstruction ability of latent inversion, we suggest timestep-aware text injection sampling. This effectively retains the structure of the input image by injecting the source text prompt in early sampling steps and then transitioning to the target prompt in subsequent sampling steps. This strategic approach seamlessly harmonizes with text optimization, facilitating complex non-rigid edits to the input without losing the original identity. We demonstrate the effectiveness of our method in terms of identity preservation, editability, and aesthetic quality through extensive experiments.

Wonjeong Ryu, Seungjun Yu, Seokha Moon et al.

End-to-End (E2E) planning has become a powerful paradigm for autonomous driving, yet current systems remain fundamentally uncertainty-blind. They assume perception outputs are fully reliable, even in ambiguous or poorly observed scenes, leaving the planner without an explicit measure of uncertainty. To address this limitation, we propose a camera-only E2E framework that estimates aleatoric uncertainty directly in BEV space and incorporates it into planning. Our method produces a dense, uncertainty-aware drivability map that captures both semantic structure and geometric layout at pixel-level resolution. To further promote safe and rule-compliant behavior, we introduce a lane-following regularization that encodes lane structure and traffic norms. This prior stabilizes trajectory planning under normal conditions while preserving the flexibility needed for maneuvers such as overtaking or lane changes. Together, these components enable robust and interpretable trajectory planning, even under challenging uncertainty conditions. Evaluated on the NAVSIM benchmark, our method achieves state-of-the-art performance, delivering substantial gains on both the challenging NAVHARD and NAVSAFE subsets. These results demonstrate that our principled aleatoric uncertainty modeling combined with driving priors significantly advances the safety and reliability of camera-only E2E autonomous driving.

Eunsu Kim, Junyeong Park, Na Min An et al.

In a globalized world, cultural elements from diverse origins frequently appear together within a single visual scene. We refer to these as culture mixing scenarios, yet how Large Vision-Language Models (LVLMs) perceive them remains underexplored. We investigate culture mixing as a critical challenge for LVLMs and examine how current models behave when cultural items from multiple regions appear together. To systematically analyze these behaviors, we construct CultureMix, a food Visual Question Answering (VQA) benchmark with 23k diffusion-generated, human-verified culture mixing images across four subtasks: (1) food-only, (2) food+food, (3) food+background, and (4) food+food+background. Evaluating 10 LVLMs, we find consistent failures to preserve individual cultural identities in mixed settings. Models show strong background reliance, with accuracy dropping 14% when cultural backgrounds are added to food-only baselines, and they produce inconsistent predictions for identical foods across different contexts. To address these limitations, we explore three robustness strategies. We find supervised fine-tuning using a diverse culture mixing dataset substantially improve model consistency and reduce background sensitivity. We call for increased attention to culture mixing scenarios as a critical step toward developing LVLMs capable of operating reliably in culturally diverse real-world environments.

Inha Kang, Eunki Kim, Wonjeong Ryu et al.

Accurate long horizon forecasting of particulate matter (PM) concentration fields is essential for operational public health decisions. However, achieving reliable forecasts remains challenging in regions with complex terrain and strong atmospheric dynamics such as East Asia. While foundation models such as Aurora offer global generality, they often miss region-specific dynamics and rely on non-real-time inputs, limiting their practical utility for localized warning systems. To address this gap, we construct and release the real-world observations and high-resolution CMAQ-OBS dataset for East Asia, reducing regional error by 59.5% and enabling real-time 48-120 hour forecasts critical for public health alerts. However, standard point-wise objectives cannot reflect asymmetric operational costs, where false alarms deteriorate public trust while missed severe events endanger populations. This cost mismatch causes SFT models to over-predict and yield high False Alarm Rates. We introduce Group-Relative Policy Optimization (GRPO) with class-wise rewards and curriculum rollout to align predictions with operational priorities. Experimental results demonstrate that our framework significantly improves the reliability of the forecast. Compared to the SFT-only baseline, our model reduces the False Alarm Rate by 47.3% while achieving a competitive F1-score, proving its effectiveness for practical, real-world air quality forecasting systems on long lead time scenarios.

Seungjun Yu, Junsung Park, Youngsun Lim et al.

We present a two-phase vision-language QA system for autonomous driving that answers high-level perception, prediction, and planning questions. In Phase-1, a large multimodal LLM (Qwen2.5-VL-32B) is conditioned on six-camera inputs, a short temporal window of history, and a chain-of-thought prompt with few-shot exemplars. A self-consistency ensemble (multiple sampled reasoning chains) further improves answer reliability. In Phase-2, we augment the prompt with nuScenes scene metadata (object annotations, ego-vehicle state, etc.) and category-specific question instructions (separate prompts for perception, prediction, planning tasks). In experiments on a driving QA benchmark, our approach significantly outperforms the baseline Qwen2.5 models. For example, using 5 history frames and 10-shot prompting in Phase-1 yields 65.1% overall accuracy (vs.62.61% with zero-shot); applying self-consistency raises this to 66.85%. Phase-2 achieves 67.37% overall. Notably, the system maintains 96% accuracy under severe visual corruption. These results demonstrate that carefully engineered prompts and contextual grounding can greatly enhance high-level driving QA with pretrained vision-language models.

Hojun Choi, Youngsun Lim, Jaeyo Shin et al.

Open-vocabulary object detection (OVD) seeks to recognize and localize object categories beyond those seen during training. Recent approaches typically leverage vision-language models (VLMs) to generate pseudo-labels using image-text alignment, allowing detectors to generalize to unseen classes without explicit supervision. However, these methods depend heavily on direct image-text matching, neglecting the intermediate reasoning steps essential for interpreting semantically complex scenes. This results in limited robustness when confronted with crowded or occluded visual contexts. In this paper, we introduce CoT-PL, a new framework that employs structured visual chain-of-thought (CoT) reasoning into the pseudo-labeling process. CoT-PL decomposes object understanding into three interpretable steps: (1) region perception even for unseen objects, (2) category recognition via zero-shot reasoning, and (3) background grounding to separate semantically complex objects. Crucially, the third step naturally motivates our contrastive background learning (CBL) that uses the pre-computed background cues as negatives to promote feature disentanglement between objects and background. In this way, CoT reasoning and CBL form an integrated pipeline tailored to robust pseudo-labeling in crowded or occluded scenes. Notably, in these two settings, our novel-class pseudo-label quality achieves relative improvements of 103.4% and 168.4% over the best prior, respectively. Our extensive experiments demonstrate that CoT-PL achieves +7.7 AP50 on open-vocabulary COCO and +2.9 mask AP on LVIS for novel classes, setting a new state of the art.

Inha Kang, Youngsun Lim, Seonho Lee et al.

State-of-the-art vision-language models (VLMs) suffer from a critical failure in understanding negation, often referred to as affirmative bias. This limitation is particularly severe in described object detection (DOD) tasks. To address this, we propose two primary contributions: (1) a new dataset pipeline and (2) a novel, lightweight adaptation recipe. First, we introduce CoVAND, a dataset constructed with a systematic chain-of-thought (CoT) and VQA-based pipeline to generate high-quality, instance-grounded negation data. Second, we propose NegToMe, a novel text token merging module that directly tackles the architectural cause of affirmative bias. NegToMe fundamentally addresses the structural loss of negation cues in tokenization, grouping them with attributes into coherent semantic phrases. It maintains correct polarity at the input level, enabling robust negation understanding even with limited data. For instance, to prevent a model from treating the fragmented tokens "not" and "girl" as simply "girl", NegToMe binds them into a single token whose meaning is correctly distinguished from that of "girl" alone. This module is integrated with a parameter-efficient and strategic LoRA fine-tuning approach. Our method significantly improves performance on challenging negation benchmarks with a lowered false positive rate, boosting NMS-AP by up to +10.8 points on OVDEval and demonstrating generalization to SoTA VLMs. This work marks a crucial step forward in addressing negation understanding for real-world detection applications.

Eunki Kim, Na Min An, James Thorne et al.

Evaluating image-text alignment while reflecting human preferences across multiple aspects is a significant issue for the development of reliable vision-language applications. It becomes especially crucial in real-world scenarios where multiple valid descriptions exist depending on contexts or user needs. However, research progress is hindered by the lack of comprehensive benchmarks and existing evaluation predictors lacking at least one of these key properties: (1) Alignment with human judgments, (2) Long-sequence processing, (3) Inference efficiency, and (4) Applicability to multi-objective scoring. To address these challenges, we propose a plug-and-play architecture to build a robust predictor, MULTI-TAP (Multi-Objective Task-Aware Predictor), capable of both multi and single-objective scoring. MULTI-TAP can produce a single overall score, utilizing a reward head built on top of a large vision-language model (LVLMs). We show that MULTI-TAP is robust in terms of application to different LVLM architectures, achieving significantly higher performance than existing metrics and even on par with the GPT-4o-based predictor, G-VEval, with a smaller size (7-8B). By training a lightweight ridge regression layer on the frozen hidden states of a pre-trained LVLM, MULTI-TAP can produce fine-grained scores for multiple human-interpretable objectives. MULTI-TAP performs better than VisionREWARD, a high-performing multi-objective reward model, in both performance and efficiency on multi-objective benchmarks and our newly released text-image-to-text dataset, EYE4ALL. Our new dataset, consisting of chosen/rejected human preferences (EYE4ALLPref) and human-annotated fine-grained scores across seven dimensions (EYE4ALLMulti), can serve as a foundation for developing more accessible AI systems by capturing the underlying preferences of users, including blind and low-vision (BLV) individuals.

Na Min An, Inha Kang, Minhyun Lee et al.

Spatial grounding is crucial for referring image segmentation (RIS), where the goal of the task is to localize an object described by language. Current foundational vision-language models (VLMs), such as CLIP, excel at aligning images and text but struggle with understanding spatial relationships. Within the language stream, most existing methods often focus on the primary noun phrase when extracting local text features, undermining contextual tokens. Within the vision stream, CLIP generates similar features for images with different spatial layouts, resulting in limited sensitivity to spatial structure. To address these limitations, we propose \textsc{CoPatch}, a zero-shot RIS framework that leverages internal model components to enhance spatial representations in both text and image modalities. For language, \textsc{CoPatch} constructs hybrid text features by incorporating context tokens carrying spatial cues. For vision, it extracts patch-level image features using our novel path discovered from intermediate layers, where spatial structure is better preserved. These enhanced features are fused into a clustered image-text similarity map, \texttt{CoMap}, enabling precise mask selection. As a result, \textsc{CoPatch} significantly improves spatial grounding in zero-shot RIS across RefCOCO, RefCOCO+, RefCOCOg, and PhraseCut (+ 2--7 mIoU) without requiring any additional training. Our findings underscore the importance of recovering and leveraging the untapped spatial knowledge inherently embedded in VLMs, thereby paving the way for opportunities in zero-shot RIS.

Raehyuk Jung, Seungjun Yu, Hyunjung Shim

Vision-Language Models (VLMs) combine a vision encoder and a large language model (LLM) through alignment training, showing strong performance on multimodal tasks. A central component in this architecture is the projection layer, which maps visual features into the LLM's embedding space. Despite its importance, its ability to generalize to unseen visual concepts has not been systematically evaluated. To address this, we propose a benchmark for evaluating projection-layer generalization. We adapt object detection datasets (rich in fine-grained annotations) into a prompting format and design train/test splits with disjoint label sets, enabling precise control over seen and unseen concept separation. Experimental results show that the projection layer retains about 79 to 88 percent of the performance on unseen classes compared to seen ones across various settings, suggesting a non-trivial level of generalization even without explicit alignment supervision on those concepts. We further analyze this behavior through a mechanistic interpretability lens. Our findings indicate that the feed-forward network in the projection layer functions like a key-value memory, processing seen and unseen tokens in similar ways. This study introduces a new evaluation framework for alignment generalization and highlights the potential for efficient VLM training with limited aligned data.

Jiho Choi, Seojeong Park, Seongjong Song et al.

We present a novel training-free framework, \textit{PosterForest}, for automated scientific poster generation. Unlike prior approaches, which largely neglect the hierarchical structure of scientific documents and the semantic integration of textual and visual elements, our method addresses both challenges directly. We introduce the \textit{Poster Tree}, a hierarchical intermediate representation that jointly encodes document structure and visual-textual relationships at multiple levels. Our framework employs a multi-agent collaboration strategy, where agents specializing in content summarization and layout planning iteratively coordinate and provide mutual feedback. This approach enables the joint optimization of logical consistency, content fidelity, and visual coherence. Extensive experiments on multiple academic domains show that our method outperforms existing baselines in both qualitative and quantitative evaluations. The resulting posters achieve quality closest to expert-designed ground truth and deliver superior information preservation, structural clarity, and user preference.

Junyong Kang, Seohyun Lim, Kyungjune Baek et al.

Aligning text-to-image (T2I) diffusion models with human preferences has emerged as a critical research challenge. While recent advances in this area have extended preference optimization techniques from large language models (LLMs) to the diffusion setting, they often struggle with limited exploration. In this work, we propose a novel and orthogonal approach to enhancing diffusion-based preference optimization. First, we introduce a stable reference model update strategy that relaxes the frozen reference model, encouraging exploration while maintaining a stable optimization anchor through reference model regularization. Second, we present a timestep-aware training strategy that mitigates the reward scale imbalance problem across timesteps. Our method can be integrated into various preference optimization algorithms. Experimental results show that our approach improves the performance of state-of-the-art methods on human preference evaluation benchmarks.

Hojun Choi, Junsuk Choe, Hyunjung Shim

Existing open-vocabulary object detection (OVD) develops methods for testing unseen categories by aligning object region embeddings with corresponding VLM features. A recent study leverages the idea that VLMs implicitly learn compositional structures of semantic concepts within the image. Instead of using an individual region embedding, it utilizes a bag of region embeddings as a new representation to incorporate compositional structures into the OVD task. However, this approach often fails to capture the contextual concepts of each region, leading to noisy compositional structures. This results in only marginal performance improvements and reduced efficiency. To address this, we propose a novel concept-based alignment method that samples a more powerful and efficient compositional structure. Our approach groups contextually related ``concepts'' into a bag and adjusts the scale of concepts within the bag for more effective embedding alignment. Combined with Faster R-CNN, our method achieves improvements of 2.6 box AP50 and 0.5 mask AP over prior work on novel categories in the open-vocabulary COCO and LVIS benchmarks. Furthermore, our method reduces CLIP computation in FLOPs by 80.3% compared to previous research, significantly enhancing efficiency. Experimental results demonstrate that the proposed method outperforms previous state-of-the-art models on the OVD datasets.

Junsung Park, Hyunjung Shim

Weakly Supervised Semantic Segmentation (WSSS) employs weak supervision, such as image-level labels, to train the segmentation model. Despite the impressive achievement in recent WSSS methods, we identify that introducing weak labels with high mean Intersection of Union (mIoU) does not guarantee high segmentation performance. Existing studies have emphasized the importance of prioritizing precision and reducing noise to improve overall performance. In the same vein, we propose ORANDNet, an advanced ensemble approach tailored for WSSS. ORANDNet combines Class Activation Maps (CAMs) from two different classifiers to increase the precision of pseudo-masks (PMs). To further mitigate small noise in the PMs, we incorporate curriculum learning. This involves training the segmentation model initially with pairs of smaller-sized images and corresponding PMs, gradually transitioning to the original-sized pairs. By combining the original CAMs of ResNet-50 and ViT, we significantly improve the segmentation performance over the single-best model and the naive ensemble model, respectively. We further extend our ensemble method to CAMs from AMN (ResNet-like) and MCTformer (ViT-like) models, achieving performance benefits in advanced WSSS models. It highlights the potential of our ORANDNet as a final add-on module for WSSS models.

Jiho Choi, Seonho Lee, Seungho Lee et al.

Open-vocabulary part segmentation (OVPS) is an emerging research area focused on segmenting fine-grained entities using diverse and previously unseen vocabularies. Our study highlights the inherent complexities of part segmentation due to intricate boundaries and diverse granularity, reflecting the knowledge-based nature of part identification. To address these challenges, we propose PartCLIPSeg, a novel framework utilizing generalized parts and object-level contexts to mitigate the lack of generalization in fine-grained parts. PartCLIPSeg integrates competitive part relationships and attention control, alleviating ambiguous boundaries and underrepresented parts. Experimental results demonstrate that PartCLIPSeg outperforms existing state-of-the-art OVPS methods, offering refined segmentation and an advanced understanding of part relationships within images. Through extensive experiments, our model demonstrated a significant improvement over the state-of-the-art models on the Pascal-Part-116, ADE20K-Part-234, and PartImageNet datasets.

Seungho Lee, Seoungyoon Kang, Hyunjung Shim

This study demonstrates a cost-effective approach to semantic segmentation using self-supervised vision transformers (SSVT). By freezing the SSVT backbone and training a lightweight segmentation head, our approach effectively utilizes imperfect labels, thereby improving robustness to label imperfections. Empirical experiments show significant performance improvements over existing methods for various annotation types, including scribble, point-level, and image-level labels. The research highlights the effectiveness of self-supervised vision transformers in dealing with imperfect labels, providing a practical and efficient solution for semantic segmentation while reducing annotation costs. Through extensive experiments, we confirm that our method outperforms baseline models for all types of imperfect labels. Especially under the zero-shot vision-language-model-based label, our model exhibits 11.5\%p performance gain compared to the baseline.

Minhyun Lee, Dongseob Kim, Hyunjung Shim

Weakly-supervised semantic segmentation (WSSS) has recently gained much attention for its promise to train segmentation models only with image-level labels. Existing WSSS methods commonly argue that the sparse coverage of CAM incurs the performance bottleneck of WSSS. This paper provides analytical and empirical evidence that the actual bottleneck may not be sparse coverage but a global thresholding scheme applied after CAM. Then, we show that this issue can be mitigated by satisfying two conditions; 1) reducing the imbalance in the foreground activation and 2) increasing the gap between the foreground and the background activation. Based on these findings, we propose a novel activation manipulation network with a per-pixel classification loss and a label conditioning module. Per-pixel classification naturally induces two-level activation in activation maps, which can penalize the most discriminative parts, promote the less discriminative parts, and deactivate the background regions. Label conditioning imposes that the output label of pseudo-masks should be any of true image-level labels; it penalizes the wrong activation assigned to non-target classes. Based on extensive analysis and evaluations, we demonstrate that each component helps produce accurate pseudo-masks, achieving the robustness against the choice of the global threshold. Finally, our model achieves state-of-the-art records on both PASCAL VOC 2012 and MS COCO 2014 datasets.

Minjin Choi, Jinhong Kim, Joonsek Lee et al.

Session-based recommendation (SR) predicts the next items from a sequence of previous items consumed by an anonymous user. Most existing SR models focus only on modeling intra-session characteristics but pay less attention to inter-session relationships of items, which has the potential to improve accuracy. Another critical aspect of recommender systems is computational efficiency and scalability, considering practical feasibility in commercial applications. To account for both accuracy and scalability, we propose a novel session-based recommendation with a random walk, namely S-Walk. Precisely, S-Walk effectively captures intra- and inter-session correlations by handling high-order relationships among items using random walks with restart (RWR). By adopting linear models with closed-form solutions for transition and teleportation matrices that constitute RWR, S-Walk is highly efficient and scalable. Extensive experiments demonstrate that S-Walk achieves comparable or state-of-the-art performance in various metrics on four benchmark datasets. Moreover, the model learned by S-Walk can be highly compressed without sacrificing accuracy, conducting two or more orders of magnitude faster inference than existing DNN-based models, making it suitable for large-scale commercial systems.

Seungho Lee, Minhyun Lee, Jongwuk Lee et al.

Existing studies in weakly-supervised semantic segmentation (WSSS) using image-level weak supervision have several limitations: sparse object coverage, inaccurate object boundaries, and co-occurring pixels from non-target objects. To overcome these challenges, we propose a novel framework, namely Explicit Pseudo-pixel Supervision (EPS), which learns from pixel-level feedback by combining two weak supervisions; the image-level label provides the object identity via the localization map and the saliency map from the off-the-shelf saliency detection model offers rich boundaries. We devise a joint training strategy to fully utilize the complementary relationship between both information. Our method can obtain accurate object boundaries and discard co-occurring pixels, thereby significantly improving the quality of pseudo-masks. Experimental results show that the proposed method remarkably outperforms existing methods by resolving key challenges of WSSS and achieves the new state-of-the-art performance on both PASCAL VOC 2012 and MS COCO 2014 datasets.