Bo Ren

Abdelrahman Abouelenin, Atabak Ashfaq, Adam Atkinson et al. · microsoft-research

We introduce Phi-4-Mini and Phi-4-Multimodal, compact yet highly capable language and multimodal models. Phi-4-Mini is a 3.8-billion-parameter language model trained on high-quality web and synthetic data, significantly outperforming recent open-source models of similar size and matching the performance of models twice its size on math and coding tasks requiring complex reasoning. This achievement is driven by a carefully curated synthetic data recipe emphasizing high-quality math and coding datasets. Compared to its predecessor, Phi-3.5-Mini, Phi-4-Mini features an expanded vocabulary size of 200K tokens to better support multilingual applications, as well as group query attention for more efficient long-sequence generation. Phi-4-Multimodal is a multimodal model that integrates text, vision, and speech/audio input modalities into a single model. Its novel modality extension approach leverages LoRA adapters and modality-specific routers to allow multiple inference modes combining various modalities without interference. For example, it now ranks first in the OpenASR leaderboard to date, although the LoRA component of the speech/audio modality has just 460 million parameters. Phi-4-Multimodal supports scenarios involving (vision + language), (vision + speech), and (speech/audio) inputs, outperforming larger vision-language and speech-language models on a wide range of tasks. Additionally, we experiment to further train Phi-4-Mini to enhance its reasoning capabilities. Despite its compact 3.8-billion-parameter size, this experimental version achieves reasoning performance on par with or surpassing significantly larger models, including DeepSeek-R1-Distill-Qwen-7B and DeepSeek-R1-Distill-Llama-8B.

Hao Wang, Junchao Liao, Tianheng Cheng et al.

Recently, the semantics of scene text has been proven to be essential in fine-grained image classification. However, the existing methods mainly exploit the literal meaning of scene text for fine-grained recognition, which might be irrelevant when it is not significantly related to objects/scenes. We propose an end-to-end trainable network that mines implicit contextual knowledge behind scene text image and enhance the semantics and correlation to fine-tune the image representation. Unlike the existing methods, our model integrates three modalities: visual feature extraction, text semantics extraction, and correlating background knowledge to fine-grained image classification. Specifically, we employ KnowBert to retrieve relevant knowledge for semantic representation and combine it with image features for fine-grained classification. Experiments on two benchmark datasets, Con-Text, and Drink Bottle, show that our method outperforms the state-of-the-art by 3.72\% mAP and 5.39\% mAP, respectively. To further validate the effectiveness of the proposed method, we create a new dataset on crowd activity recognition for the evaluation. The source code and new dataset of this work are available at https://github.com/lanfeng4659/KnowledgeMiningWithSceneText.

Sunan He, Taian Guo, Tao Dai et al.

Real-world recognition system often encounters the challenge of unseen labels. To identify such unseen labels, multi-label zero-shot learning (ML-ZSL) focuses on transferring knowledge by a pre-trained textual label embedding (e.g., GloVe). However, such methods only exploit single-modal knowledge from a language model, while ignoring the rich semantic information inherent in image-text pairs. Instead, recently developed open-vocabulary (OV) based methods succeed in exploiting such information of image-text pairs in object detection, and achieve impressive performance. Inspired by the success of OV-based methods, we propose a novel open-vocabulary framework, named multi-modal knowledge transfer (MKT), for multi-label classification. Specifically, our method exploits multi-modal knowledge of image-text pairs based on a vision and language pre-training (VLP) model. To facilitate transferring the image-text matching ability of VLP model, knowledge distillation is employed to guarantee the consistency of image and label embeddings, along with prompt tuning to further update the label embeddings. To further enable the recognition of multiple objects, a simple but effective two-stream module is developed to capture both local and global features. Extensive experimental results show that our method significantly outperforms state-of-the-art methods on public benchmark datasets. The source code is available at https://github.com/sunanhe/MKT.

Xiujun Shu, Wei Wen, Haoqian Wu et al.

Text-based person retrieval aims to find the query person based on a textual description. The key is to learn a common latent space mapping between visual-textual modalities. To achieve this goal, existing works employ segmentation to obtain explicitly cross-modal alignments or utilize attention to explore salient alignments. These methods have two shortcomings: 1) Labeling cross-modal alignments are time-consuming. 2) Attention methods can explore salient cross-modal alignments but may ignore some subtle and valuable pairs. To relieve these issues, we introduce an Implicit Visual-Textual (IVT) framework for text-based person retrieval. Different from previous models, IVT utilizes a single network to learn representation for both modalities, which contributes to the visual-textual interaction. To explore the fine-grained alignment, we further propose two implicit semantic alignment paradigms: multi-level alignment (MLA) and bidirectional mask modeling (BMM). The MLA module explores finer matching at sentence, phrase, and word levels, while the BMM module aims to mine \textbf{more} semantic alignments between visual and textual modalities. Extensive experiments are carried out to evaluate the proposed IVT on public datasets, i.e., CUHK-PEDES, RSTPReID, and ICFG-PEDES. Even without explicit body part alignment, our approach still achieves state-of-the-art performance. Code is available at: https://github.com/TencentYoutuResearch/PersonRetrieval-IVT.

Wenwen Yu, Yuliang Liu, Wei Hua et al.

The recent large-scale Contrastive Language-Image Pretraining (CLIP) model has shown great potential in various downstream tasks via leveraging the pretrained vision and language knowledge. Scene text, which contains rich textual and visual information, has an inherent connection with a model like CLIP. Recently, pretraining approaches based on vision language models have made effective progresses in the field of text detection. In contrast to these works, this paper proposes a new method, termed TCM, focusing on Turning the CLIP Model directly for text detection without pretraining process. We demonstrate the advantages of the proposed TCM as follows: (1) The underlying principle of our framework can be applied to improve existing scene text detector. (2) It facilitates the few-shot training capability of existing methods, e.g., by using 10% of labeled data, we significantly improve the performance of the baseline method with an average of 22% in terms of the F-measure on 4 benchmarks. (3) By turning the CLIP model into existing scene text detection methods, we further achieve promising domain adaptation ability. The code will be publicly released at https://github.com/wenwenyu/TCM.

Yulei Qin, Xingyu Chen, Chao Chen et al.

Recently, webly supervised learning (WSL) has been studied to leverage numerous and accessible data from the Internet. Most existing methods focus on learning noise-robust models from web images while neglecting the performance drop caused by the differences between web domain and real-world domain. However, only by tackling the performance gap above can we fully exploit the practical value of web datasets. To this end, we propose a Few-shot guided Prototypical (FoPro) representation learning method, which only needs a few labeled examples from reality and can significantly improve the performance in the real-world domain. Specifically, we initialize each class center with few-shot real-world data as the ``realistic" prototype. Then, the intra-class distance between web instances and ``realistic" prototypes is narrowed by contrastive learning. Finally, we measure image-prototype distance with a learnable metric. Prototypes are polished by adjacent high-quality web images and involved in removing distant out-of-distribution samples. In experiments, FoPro is trained on web datasets with a few real-world examples guided and evaluated on real-world datasets. Our method achieves the state-of-the-art performance on three fine-grained datasets and two large-scale datasets. Compared with existing WSL methods under the same few-shot settings, FoPro still excels in real-world generalization. Code is available at https://github.com/yuleiqin/fopro.

Jian Wu, Yashesh Gaur, Zhuo Chen et al.

Large language models (LLMs) have achieved remarkable success in the field of natural language processing, enabling better human-computer interaction using natural language. However, the seamless integration of speech signals into LLMs has not been explored well. The "decoder-only" architecture has also not been well studied for speech processing tasks. In this research, we introduce Speech-LLaMA, a novel approach that effectively incorporates acoustic information into text-based large language models. Our method leverages Connectionist Temporal Classification and a simple audio encoder to map the compressed acoustic features to the continuous semantic space of the LLM. In addition, we further probe the decoder-only architecture for speech-to-text tasks by training a smaller scale randomly initialized speech-LLaMA model from speech-text paired data alone. We conduct experiments on multilingual speech-to-text translation tasks and demonstrate a significant improvement over strong baselines, highlighting the potential advantages of decoder-only models for speech-to-text conversion.

Yuan Liang, Zhuoxuan Jiang, Di Yin et al.

In document-level event extraction (DEE) task, event arguments always scatter across sentences (across-sentence issue) and multiple events may lie in one document (multi-event issue). In this paper, we argue that the relation information of event arguments is of great significance for addressing the above two issues, and propose a new DEE framework which can model the relation dependencies, called Relation-augmented Document-level Event Extraction (ReDEE). More specifically, this framework features a novel and tailored transformer, named as Relation-augmented Attention Transformer (RAAT). RAAT is scalable to capture multi-scale and multi-amount argument relations. To further leverage relation information, we introduce a separate event relation prediction task and adopt multi-task learning method to explicitly enhance event extraction performance. Extensive experiments demonstrate the effectiveness of the proposed method, which can achieve state-of-the-art performance on two public datasets. Our code is available at https://github. com/TencentYoutuResearch/RAAT.

Jiaxiong Qiu, Peng-Tao Jiang, Yifan Zhu et al.

Neural implicit methods have achieved high-quality 3D object surfaces under slight specular highlights. However, high specular reflections (HSR) often appear in front of target objects when we capture them through glasses. The complex ambiguity in these scenes violates the multi-view consistency, then makes it challenging for recent methods to reconstruct target objects correctly. To remedy this issue, we present a novel surface reconstruction framework, NeuS-HSR, based on implicit neural rendering. In NeuS-HSR, the object surface is parameterized as an implicit signed distance function (SDF). To reduce the interference of HSR, we propose decomposing the rendered image into two appearances: the target object and the auxiliary plane. We design a novel auxiliary plane module by combining physical assumptions and neural networks to generate the auxiliary plane appearance. Extensive experiments on synthetic and real-world datasets demonstrate that NeuS-HSR outperforms state-of-the-art approaches for accurate and robust target surface reconstruction against HSR. Code is available at https://github.com/JiaxiongQ/NeuS-HSR.

Peixian Chen, Kekai Sheng, Mengdan Zhang et al.

Open-vocabulary object detection (OVD) aims to scale up vocabulary size to detect objects of novel categories beyond the training vocabulary. Recent work resorts to the rich knowledge in pre-trained vision-language models. However, existing methods are ineffective in proposal-level vision-language alignment. Meanwhile, the models usually suffer from confidence bias toward base categories and perform worse on novel ones. To overcome the challenges, we present MEDet, a novel and effective OVD framework with proposal mining and prediction equalization. First, we design an online proposal mining to refine the inherited vision-semantic knowledge from coarse to fine, allowing for proposal-level detection-oriented feature alignment. Second, based on causal inference theory, we introduce a class-wise backdoor adjustment to reinforce the predictions on novel categories to improve the overall OVD performance. Extensive experiments on COCO and LVIS benchmarks verify the superiority of MEDet over the competing approaches in detecting objects of novel categories, e.g., 32.6% AP50 on COCO and 22.4% mask mAP on LVIS.

Hao Liu, Xinghua Jiang, Xin Li et al. · bytedance, tencent-ai

The self-supervised Masked Image Modeling (MIM) schema, following "mask-and-reconstruct" pipeline of recovering contents from masked image, has recently captured the increasing interest in the multimedia community, owing to the excellent ability of learning visual representation from unlabeled data. Aiming at learning representations with high semantics abstracted, a group of works attempts to reconstruct non-semantic pixels with large-ratio masking strategy, which may suffer from "over-smoothing" problem, while others directly infuse semantics into targets in off-line way requiring extra data. Different from them, we shift the perspective to the Fourier domain which naturally has global perspective and present a new Masked Image Modeling (MIM), termed Geminated Gestalt Autoencoder (Ge$^2$-AE) for visual pre-training. Specifically, we equip our model with geminated decoders in charge of reconstructing image contents from both pixel and frequency space, where each other serves as not only the complementation but also the reciprocal constraints. Through this way, more robust representations can be learned in the pre-trained encoders, of which the effectiveness is confirmed by the juxtaposing experimental results on downstream recognition tasks. We also conduct several quantitative and qualitative experiments to investigate the learning behavior of our method. To our best knowledge, this is the first MIM work to solve the visual pre-training through the lens of frequency domain.

Bei Gan, Xiujun Shu, Ruizhi Qiao et al.

Movie highlights stand out of the screenplay for efficient browsing and play a crucial role on social media platforms. Based on existing efforts, this work has two observations: (1) For different annotators, labeling highlight has uncertainty, which leads to inaccurate and time-consuming annotations. (2) Besides previous supervised or unsupervised settings, some existing video corpora can be useful, e.g., trailers, but they are often noisy and incomplete to cover the full highlights. In this work, we study a more practical and promising setting, i.e., reformulating highlight detection as "learning with noisy labels". This setting does not require time-consuming manual annotations and can fully utilize existing abundant video corpora. First, based on movie trailers, we leverage scene segmentation to obtain complete shots, which are regarded as noisy labels. Then, we propose a Collaborative noisy Label Cleaner (CLC) framework to learn from noisy highlight moments. CLC consists of two modules: augmented cross-propagation (ACP) and multi-modality cleaning (MMC). The former aims to exploit the closely related audio-visual signals and fuse them to learn unified multi-modal representations. The latter aims to achieve cleaner highlight labels by observing the changes in losses among different modalities. To verify the effectiveness of CLC, we further collect a large-scale highlight dataset named MovieLights. Comprehensive experiments on MovieLights and YouTube Highlights datasets demonstrate the effectiveness of our approach. Code has been made available at: https://github.com/TencentYoutuResearch/HighlightDetection-CLC

Zhihao Li, Biao Hou, Siteng Ma et al.

To overcome the inherent domain gap between remote sensing (RS) images and natural images, some self-supervised representation learning methods have made promising progress. However, they have overlooked the diverse angles present in RS objects. This paper proposes the Masked Angle-Aware Autoencoder (MA3E) to perceive and learn angles during pre-training. We design a \textit{scaling center crop} operation to create the rotated crop with random orientation on each original image, introducing the explicit angle variation. MA3E inputs this composite image while reconstruct the original image, aiming to effectively learn rotation-invariant representations by restoring the angle variation introduced on the rotated crop. To avoid biases caused by directly reconstructing the rotated crop, we propose an Optimal Transport (OT) loss that automatically assigns similar original image patches to each rotated crop patch for reconstruction. MA3E demonstrates more competitive performance than existing pre-training methods on seven different RS image datasets in three downstream tasks.

Xin Li, Yan Zheng, Yiqing Hu et al. · tencent-ai

Relational understanding is critical for a number of visually-rich documents (VRDs) understanding tasks. Through multi-modal pre-training, recent studies provide comprehensive contextual representations and exploit them as prior knowledge for downstream tasks. In spite of their impressive results, we observe that the widespread relational hints (e.g., relation of key/value fields on receipts) built upon contextual knowledge are not excavated yet. To mitigate this gap, we propose DocReL, a Document Relational Representation Learning framework. The major challenge of DocReL roots in the variety of relations. From the simplest pairwise relation to the complex global structure, it is infeasible to conduct supervised training due to the definition of relation varies and even conflicts in different tasks. To deal with the unpredictable definition of relations, we propose a novel contrastive learning task named Relational Consistency Modeling (RCM), which harnesses the fact that existing relations should be consistent in differently augmented positive views. RCM provides relational representations which are more compatible to the urgent need of downstream tasks, even without any knowledge about the exact definition of relation. DocReL achieves better performance on a wide variety of VRD relational understanding tasks, including table structure recognition, key information extraction and reading order detection.

Haoqian Wu, Keyu Chen, Yanan Luo et al.

A long-term video, such as a movie or TV show, is composed of various scenes, each of which represents a series of shots sharing the same semantic story. Spotting the correct scene boundary from the long-term video is a challenging task, since a model must understand the storyline of the video to figure out where a scene starts and ends. To this end, we propose an effective Self-Supervised Learning (SSL) framework to learn better shot representations from unlabeled long-term videos. More specifically, we present an SSL scheme to achieve scene consistency, while exploring considerable data augmentation and shuffling methods to boost the model generalizability. Instead of explicitly learning the scene boundary features as in the previous methods, we introduce a vanilla temporal model with less inductive bias to verify the quality of the shot features. Our method achieves the state-of-the-art performance on the task of Video Scene Segmentation. Additionally, we suggest a more fair and reasonable benchmark to evaluate the performance of Video Scene Segmentation methods. The code is made available.

Jiquan Li, Junliang Guo, Yongxin Zhu et al.

The task of Grammatical Error Correction (GEC) has received remarkable attention with wide applications in Natural Language Processing (NLP) in recent years. While one of the key principles of GEC is to keep the correct parts unchanged and avoid over-correction, previous sequence-to-sequence (seq2seq) models generate results from scratch, which are not guaranteed to follow the original sentence structure and may suffer from the over-correction problem. In the meantime, the recently proposed sequence tagging models can overcome the over-correction problem by only generating edit operations, but are conditioned on human designed language-specific tagging labels. In this paper, we combine the pros and alleviate the cons of both models by proposing a novel Sequence-to-Action~(S2A) module. The S2A module jointly takes the source and target sentences as input, and is able to automatically generate a token-level action sequence before predicting each token, where each action is generated from three choices named SKIP, COPY and GENerate. Then the actions are fused with the basic seq2seq framework to provide final predictions. We conduct experiments on the benchmark datasets of both English and Chinese GEC tasks. Our model consistently outperforms the seq2seq baselines, while being able to significantly alleviate the over-correction problem as well as holding better generality and diversity in the generation results compared to the sequence tagging models.

Sunan He, Taian Guo, Tao Dai et al.

Image and language modeling is of crucial importance for vision-language pre-training (VLP), which aims to learn multi-modal representations from large-scale paired image-text data. However, we observe that most existing VLP methods focus on modeling the interactions between image and text features while neglecting the information disparity between image and text, thus suffering from focal bias. To address this problem, we propose a vision-language masked autoencoder framework (VLMAE). VLMAE employs visual generative learning, facilitating the model to acquire fine-grained and unbiased features. Unlike the previous works, VLMAE pays attention to almost all critical patches in an image, providing more comprehensive understanding. Extensive experiments demonstrate that VLMAE achieves better performance in various vision-language downstream tasks, including visual question answering, image-text retrieval and visual grounding, even with up to 20% pre-training speedup.

Ye Liu, Changchong Lu, Chen Lin et al.

The extraction of text information in videos serves as a critical step towards semantic understanding of videos. It usually involved in two steps: (1) text recognition and (2) text classification. To localize texts in videos, we can resort to large numbers of text recognition methods based on OCR technology. However, to our knowledge, there is no existing work focused on the second step of video text classification, which will limit the guidance to downstream tasks such as video indexing and browsing. In this paper, we are the first to address this new task of video text classification by fusing multimodal information to deal with the challenging scenario where different types of video texts may be confused with various colors, unknown fonts and complex layouts. In addition, we tailor a specific module called CorrelationNet to reinforce feature representation by explicitly extracting layout information. Furthermore, contrastive learning is utilized to explore inherent connections between samples using plentiful unlabeled videos. Finally, we construct a new well-defined industrial dataset from the news domain, called TI-News, which is dedicated to building and evaluating video text recognition and classification applications. Extensive experiments on TI-News demonstrate the effectiveness of our method.

Haoyu Cao, Jiefeng Ma, Antai Guo et al.

Document Information Extraction (DIE) has attracted increasing attention due to its various advanced applications in the real world. Although recent literature has already achieved competitive results, these approaches usually fail when dealing with complex documents with noisy OCR results or mutative layouts. This paper proposes Generative Multi-modal Network (GMN) for real-world scenarios to address these problems, which is a robust multi-modal generation method without predefined label categories. With the carefully designed spatial encoder and modal-aware mask module, GMN can deal with complex documents that are hard to serialized into sequential order. Moreover, GMN tolerates errors in OCR results and requires no character-level annotation, which is vital because fine-grained annotation of numerous documents is laborious and even requires annotators with specialized domain knowledge. Extensive experiments show that GMN achieves new state-of-the-art performance on several public DIE datasets and surpasses other methods by a large margin, especially in realistic scenes.

Ye Liu, Lingfeng Qiao, Di Yin et al.

Scene segmentation and classification (SSC) serve as a critical step towards the field of video structuring analysis. Intuitively, jointly learning of these two tasks can promote each other by sharing common information. However, scene segmentation concerns more on the local difference between adjacent shots while classification needs the global representation of scene segments, which probably leads to the model dominated by one of the two tasks in the training phase. In this paper, from an alternate perspective to overcome the above challenges, we unite these two tasks into one task by a new form of predicting shots link: a link connects two adjacent shots, indicating that they belong to the same scene or category. To the end, we propose a general One Stage Multimodal Sequential Link Framework (OS-MSL) to both distinguish and leverage the two-fold semantics by reforming the two learning tasks into a unified one. Furthermore, we tailor a specific module called DiffCorrNet to explicitly extract the information of differences and correlations among shots. Extensive experiments on a brand-new large scale dataset collected from real-world applications, and MovieScenes are conducted. Both the results demonstrate the effectiveness of our proposed method against strong baselines.

Lingfeng Qiao, Chen Wu, Ye Liu et al.

Multimodal headline utilizes both video frames and transcripts to generate the natural language title of the videos. Due to a lack of large-scale, manually annotated data, the task of annotating grounded headlines for video is labor intensive and impractical. Previous researches on pre-trained language models and video-language models have achieved significant progress in related downstream tasks. However, none of them can be directly applied to multimodal headline architecture where we need both multimodal encoder and sentence decoder. A major challenge in simply gluing language model and video-language model is the modality balance, which is aimed at combining visual-language complementary abilities. In this paper, we propose a novel approach to graft the video encoder from the pre-trained video-language model on the generative pre-trained language model. We also present a consensus fusion mechanism for the integration of different components, via inter/intra modality relation. Empirically, experiments show that the grafted model achieves strong results on a brand-new dataset collected from real-world applications.

Zhuoxuan Jiang, Lingfeng Qiao, Di Yin et al.

Recent language generative models are mostly trained on large-scale datasets, while in some real scenarios, the training datasets are often expensive to obtain and would be small-scale. In this paper we investigate the challenging task of less-data constrained generation, especially when the generated news headlines are short yet expected by readers to keep readable and informative simultaneously. We highlight the key information modeling task and propose a novel duality fine-tuning method by formally defining the probabilistic duality constraints between key information prediction and headline generation tasks. The proposed method can capture more information from limited data, build connections between separate tasks, and is suitable for less-data constrained generation tasks. Furthermore, the method can leverage various pre-trained generative regimes, e.g., autoregressive and encoder-decoder models. We conduct extensive experiments to demonstrate that our method is effective and efficient to achieve improved performance in terms of language modeling metric and informativeness correctness metric on two public datasets.

Zheng Lin, Zhao Zhang, Kang-Rui Zhang et al.

Neural style transfer (NST) can create impressive artworks by transferring reference style to content image. Current image-to-image NST methods are short of fine-grained controls, which are often demanded by artistic editing. To mitigate this limitation, we propose a drawing-like interactive style transfer (IST) method, by which users can interactively create a harmonious-style image. Our IST method can serve as a brush, dip style from anywhere, and then paint to any region of the target content image. To determine the action scope, we formulate a fluid simulation algorithm, which takes styles as pigments around the position of brush interaction, and diffusion in style or content images according to the similarity maps. Our IST method expands the creative dimension of NST. By dipping and painting, even employing one style image can produce thousands of eye-catching works. The demo video is available in supplementary files or in http://mmcheng.net/ist.

Chang Nie, Yiqing Hu, Yanqiu Qu et al.

As textual attributes like font are core design elements of document format and page style, automatic attributes recognition favor comprehensive practical applications. Existing approaches already yield satisfactory performance in differentiating disparate attributes, but they still suffer in distinguishing similar attributes with only subtle difference. Moreover, their performance drop severely in real-world scenarios where unexpected and obvious imaging distortions appear. In this paper, we aim to tackle these problems by proposing TaCo, a contrastive framework for textual attribute recognition tailored toward the most common document scenes. Specifically, TaCo leverages contrastive learning to dispel the ambiguity trap arising from vague and open-ended attributes. To realize this goal, we design the learning paradigm from three perspectives: 1) generating attribute views, 2) extracting subtle but crucial details, and 3) exploiting valued view pairs for learning, to fully unlock the pre-training potential. Extensive experiments show that TaCo surpasses the supervised counterparts and advances the state-of-the-art remarkably on multiple attribute recognition tasks. Online services of TaCo will be made available.

Jiang-Tian Zhai, Qi Zhang, Tong Wu et al.

Learning fine-grained interplay between vision and language allows to a more accurate understanding for VisionLanguage tasks. However, it remains challenging to extract key image regions according to the texts for semantic alignments. Most existing works are either limited by textagnostic and redundant regions obtained with the frozen detectors, or failing to scale further due to its heavy reliance on scarce grounding (gold) data to pre-train detectors. To solve these problems, we propose Self-Locator Aided Network (SLAN) for cross-modal understanding tasks without any extra gold data. SLAN consists of a region filter and a region adaptor to localize regions of interest conditioned on different texts. By aggregating cross-modal information, the region filter selects key regions and the region adaptor updates their coordinates with text guidance. With detailed region-word alignments, SLAN can be easily generalized to many downstream tasks. It achieves fairly competitive results on five cross-modal understanding tasks (e.g., 85.7% and 69.2% on COCO image-to-text and text-to-image retrieval, surpassing previous SOTA methods). SLAN also demonstrates strong zero-shot and fine-tuned transferability to two localization tasks.

Keqi Deng, Ruchao Fan, Bo Ren et al.

Despite extensions to speech inputs, effectively leveraging the rich knowledge and contextual understanding of large language models (LLMs) in automatic speech recognition (ASR) remains non-trivial, as the task primarily involves direct speech-to-text mapping. To address this, this paper proposes chain-of-thought ASR (CoT-ASR), which constructs a reasoning chain that enables LLMs to first analyze the input speech and generate contextual analysis, thereby fully exploiting their generative capabilities. With this contextual reasoning, CoT-ASR then performs more informed speech recognition and completes both reasoning and transcription in a single pass. Moreover, CoT-ASR naturally supports user-guided transcription: while designed to self-generate reasoning, it can also seamlessly incorporate user-provided context to guide transcription, further extending ASR functionality. To reduce the modality gap, this paper introduces a CTC-guided Modality Adapter, which uses CTC non-blank token probabilities to weight LLM embeddings, efficiently aligning speech encoder outputs with the LLM's textual latent space. Experiments show that, compared to standard LLM-based ASR, CoT-ASR achieves a relative reduction of 8.7% in word error rate (WER) and 16.9% in entity error rate (EER).

Xin Jin, Pengyi Jiao, Zheng-Peng Duan et al.

Volumetric rendering based methods, like NeRF, excel in HDR view synthesis from RAWimages, especially for nighttime scenes. While, they suffer from long training times and cannot perform real-time rendering due to dense sampling requirements. The advent of 3D Gaussian Splatting (3DGS) enables real-time rendering and faster training. However, implementing RAW image-based view synthesis directly using 3DGS is challenging due to its inherent drawbacks: 1) in nighttime scenes, extremely low SNR leads to poor structure-from-motion (SfM) estimation in distant views; 2) the limited representation capacity of spherical harmonics (SH) function is unsuitable for RAW linear color space; and 3) inaccurate scene structure hampers downstream tasks such as refocusing. To address these issues, we propose LE3D (Lighting Every darkness with 3DGS). Our method proposes Cone Scatter Initialization to enrich the estimation of SfM, and replaces SH with a Color MLP to represent the RAW linear color space. Additionally, we introduce depth distortion and near-far regularizations to improve the accuracy of scene structure for downstream tasks. These designs enable LE3D to perform real-time novel view synthesis, HDR rendering, refocusing, and tone-mapping changes. Compared to previous volumetric rendering based methods, LE3D reduces training time to 1% and improves rendering speed by up to 4,000 times for 2K resolution images in terms of FPS. Code and viewer can be found in https://github.com/Srameo/LE3D .

Xiujun Shu, Yusheng Tao, Ruizhi Qiao et al.

Person search in media has seen increasing potential in Internet applications, such as video clipping and character collection. This task is common but overlooked by previous person search works which focus on surveillance scenes. The media scenarios have some different challenges from surveillance scenes. For example, a person may change his clothes frequently. To alleviate this issue, this paper proposes a Unified Detector and Graph Network (UDGNet) for person search in media. UDGNet is the first person search framework to detect and re-identify the human body and head simultaneously. Specifically, it first builds two branches based on a unified network to detect the human body and head, then the detected body and head are used for re-identification. This dual-task approach can significantly enhance discriminative learning. To tackle the cloth-changing issue, UDGNet builds two graphs to explore reliable links among cloth-changing samples and utilizes a graph network to learn better embeddings. This design effectively enhances the robustness of person search to cloth-changing challenges. Besides, we demonstrate that UDGNet can be implemented with both anchor-based and anchor-free person search frameworks and further achieve performance improvement. This paper also contributes a large-scale dataset for Person Search in Media (PSM), which provides both body and head annotations. It is by far the largest dataset for person search in media. Experiments show that UDGNet improves the anchor-free model AlignPS by 12.1% in mAP. Meanwhile, it shows good generalization across surveillance and longterm scenarios. The dataset and code will be available at: https://github.com/shuxjweb/PSM.git.

Hao Liu, Xinghua Jiang, Xin Li et al.

Recently, Vision Transformers (ViT), with the self-attention (SA) as the de facto ingredients, have demonstrated great potential in the computer vision community. For the sake of trade-off between efficiency and performance, a group of works merely perform SA operation within local patches, whereas the global contextual information is abandoned, which would be indispensable for visual recognition tasks. To solve the issue, the subsequent global-local ViTs take a stab at marrying local SA with global one in parallel or alternative way in the model. Nevertheless, the exhaustively combined local and global context may exist redundancy for various visual data, and the receptive field within each layer is fixed. Alternatively, a more graceful way is that global and local context can adaptively contribute per se to accommodate different visual data. To achieve this goal, we in this paper propose a novel ViT architecture, termed NomMer, which can dynamically Nominate the synergistic global-local context in vision transforMer. By investigating the working pattern of our proposed NomMer, we further explore what context information is focused. Beneficial from this "dynamic nomination" mechanism, without bells and whistles, the NomMer can not only achieve 84.5% Top-1 classification accuracy on ImageNet with only 73M parameters, but also show promising performance on dense prediction tasks, i.e., object detection and semantic segmentation. The code and models will be made publicly available at https://github.com/TencentYoutuResearch/VisualRecognition-NomMer

Zhonghua Li, Biao Hou, Zitong Wu et al.

Existing anchor-base oriented object detection methods have achieved amazing results, but these methods require some manual preset boxes, which introduces additional hyperparameters and calculations. The existing anchor-free methods usually have complex architectures and are not easy to deploy. Our goal is to propose an algorithm which is simple and easy-to-deploy for aerial image detection. In this paper, we present a one-stage anchor-free rotated object detector (FCOSR) based on FCOS, which can be deployed on most platforms. The FCOSR has a simple architecture consisting of only convolution layers. Our work focuses on the label assignment strategy for the training phase. We use ellipse center sampling method to define a suitable sampling region for oriented bounding box (OBB). The fuzzy sample assignment strategy provides reasonable labels for overlapping objects. To solve the insufficient sampling problem, a multi-level sampling module is designed. These strategies allocate more appropriate labels to training samples. Our algorithm achieves 79.25, 75.41, and 90.15 mAP on DOTA1.0, DOTA1.5, and HRSC2016 datasets, respectively. FCOSR demonstrates superior performance to other methods in single-scale evaluation. We convert a lightweight FCOSR model to TensorRT format, which achieves 73.93 mAP on DOTA1.0 at a speed of 10.68 FPS on Jetson Xavier NX with single scale. The code is available at: https://github.com/lzh420202/FCOSR

Yu-Huan Wu, Yun Liu, Le Zhang et al.

Recent progress on salient object detection (SOD) mainly benefits from multi-scale learning, where the high-level and low-level features collaborate in locating salient objects and discovering fine details, respectively. However, most efforts are devoted to low-level feature learning by fusing multi-scale features or enhancing boundary representations. High-level features, which although have long proven effective for many other tasks, yet have been barely studied for SOD. In this paper, we tap into this gap and show that enhancing high- level features is essential for SOD as well. To this end, we introduce an Extremely-Downsampled Network (EDN), which employs an extreme downsampling technique to effectively learn a global view of the whole image, leading to accurate salient object localization. To accomplish better multi-level feature fusion, we construct the Scale-Correlated Pyramid Convolution (SCPC) to build an elegant decoder for recovering object details from the above extreme downsampling. Extensive experiments demonstrate that EDN achieves state-of-the-art performance with real-time speed. Our efficient EDN-Lite also achieves competitive performance with a speed of 316fps. Hence, this work is expected to spark some new thinking in SOD. Code is available at https://github.com/yuhuan-wu/EDN.

Letian Huang, Dongwei Ye, Jialin Dan et al.

The emergence of neural and Gaussian-based radiance field methods has led to considerable advancements in novel view synthesis and 3D object reconstruction. Nonetheless, specular reflection and refraction continue to pose significant challenges due to the instability and incorrect overfitting of radiance fields to high-frequency light variations. Currently, even 3D Gaussian Splatting (3D-GS), as a powerful and efficient tool, falls short in recovering transparent objects with nearby contents due to the existence of apparent secondary ray effects. To address this issue, we propose TransparentGS, a fast inverse rendering pipeline for transparent objects based on 3D-GS. The main contributions are three-fold. Firstly, an efficient representation of transparent objects, transparent Gaussian primitives, is designed to enable specular refraction through a deferred refraction strategy. Secondly, we leverage Gaussian light field probes (GaussProbe) to encode both ambient light and nearby contents in a unified framework. Thirdly, a depth-based iterative probes query (IterQuery) algorithm is proposed to reduce the parallax errors in our probe-based framework. Experiments demonstrate the speed and accuracy of our approach in recovering transparent objects from complex environments, as well as several applications in computer graphics and vision.

Haibin Wu, Yuxuan Hu, Ruchao Fan et al.

Speech language models (Speech LMs) enable end-to-end speech-text modelling within a single model, offering a promising direction for spoken dialogue systems. The choice of speech-text jointly decoding paradigm plays a critical role in performance, efficiency, and alignment quality. In this work, we systematically compare representative joint speech-text decoding strategies-including the interleaved, and parallel generation paradigms-under a controlled experimental setup using the same base language model, speech tokenizer and training data. Our results show that the interleaved approach achieves the best alignment. However it suffers from slow inference due to long token sequence length. To address this, we propose a novel early-stop interleaved (ESI) pattern that not only significantly accelerates decoding but also yields slightly better performance. Additionally, we curate high-quality question answering (QA) datasets to further improve speech QA performance.

Jianfeng Kuang, Wei Hua, Dingkang Liang et al.

Visual information extraction (VIE), which aims to simultaneously perform OCR and information extraction in a unified framework, has drawn increasing attention due to its essential role in various applications like understanding receipts, goods, and traffic signs. However, as existing benchmark datasets for VIE mainly consist of document images without the adequate diversity of layout structures, background disturbs, and entity categories, they cannot fully reveal the challenges of real-world applications. In this paper, we propose a large-scale dataset consisting of camera images for VIE, which contains not only the larger variance of layout, backgrounds, and fonts but also much more types of entities. Besides, we propose a novel framework for end-to-end VIE that combines the stages of OCR and information extraction in an end-to-end learning fashion. Different from the previous end-to-end approaches that directly adopt OCR features as the input of an information extraction module, we propose to use contrastive learning to narrow the semantic gap caused by the difference between the tasks of OCR and information extraction. We evaluate the existing end-to-end methods for VIE on the proposed dataset and observe that the performance of these methods has a distinguishable drop from SROIE (a widely used English dataset) to our proposed dataset due to the larger variance of layout and entities. These results demonstrate our dataset is more practical for promoting advanced VIE algorithms. In addition, experiments demonstrate that the proposed VIE method consistently achieves the obvious performance gains on the proposed and SROIE datasets.

Ze-Xin Yin, Peng-Yi Jiao, Jiaxiong Qiu et al.

Existing Neural Radiance Fields (NeRF) methods suffer from the existence of reflective objects, often resulting in blurry or distorted rendering. Instead of calculating a single radiance field, we propose a multi-space neural radiance field (MS-NeRF) that represents the scene using a group of feature fields in parallel sub-spaces, which leads to a better understanding of the neural network toward the existence of reflective and refractive objects. Our multi-space scheme works as an enhancement to existing NeRF methods, with only small computational overheads needed for training and inferring the extra-space outputs. We design different multi-space modules for representative MLP-based and grid-based NeRF methods, which improve Mip-NeRF 360 by 4.15 dB in PSNR with 0.5% extra parameters and further improve TensoRF by 2.71 dB with 0.046% extra parameters on reflective regions without degrading the rendering quality on other regions. We further construct a novel dataset consisting of 33 synthetic scenes and 7 real captured scenes with complex reflection and refraction, where we design complex camera paths to fully benchmark the robustness of NeRF-based methods. Extensive experiments show that our approach significantly outperforms the existing single-space NeRF methods for rendering high-quality scenes concerned with complex light paths through mirror-like objects. The source code, dataset, and results are available via our project page: https://zx-yin.github.io/msnerf/.

Helei Qiu, Biao Hou, Bo Ren et al.

Capturing the dependencies between joints is critical in skeleton-based action recognition task. Transformer shows great potential to model the correlation of important joints. However, the existing Transformer-based methods cannot capture the correlation of different joints between frames, which the correlation is very useful since different body parts (such as the arms and legs in "long jump") between adjacent frames move together. Focus on this problem, A novel spatio-temporal tuples Transformer (STTFormer) method is proposed. The skeleton sequence is divided into several parts, and several consecutive frames contained in each part are encoded. And then a spatio-temporal tuples self-attention module is proposed to capture the relationship of different joints in consecutive frames. In addition, a feature aggregation module is introduced between non-adjacent frames to enhance the ability to distinguish similar actions. Compared with the state-of-the-art methods, our method achieves better performance on two large-scale datasets.

Hao Liu, Xin Li, Bing Liu et al.

Recently, table structure recognition has achieved impressive progress with the help of deep graph models. Most of them exploit single visual cues of tabular elements or simply combine visual cues with other modalities via early fusion to reason their graph relationships. However, neither early fusion nor individually reasoning in terms of multiple modalities can be appropriate for all varieties of table structures with great diversity. Instead, different modalities are expected to collaborate with each other in different patterns for different table cases. In the community, the importance of intra-inter modality interactions for table structure reasoning is still unexplored. In this paper, we define it as heterogeneous table structure recognition (Hetero-TSR) problem. With the aim of filling this gap, we present a novel Neural Collaborative Graph Machines (NCGM) equipped with stacked collaborative blocks, which alternatively extracts intra-modality context and models inter-modality interactions in a hierarchical way. It can represent the intra-inter modality relationships of tabular elements more robustly, which significantly improves the recognition performance. We also show that the proposed NCGM can modulate collaborative pattern of different modalities conditioned on the context of intra-modality cues, which is vital for diversified table cases. Experimental results on benchmarks demonstrate our proposed NCGM achieves state-of-the-art performance and beats other contemporary methods by a large margin especially under challenging scenarios.

Hao Liu, Antai Guo, Deqiang Jiang et al.

Recently, a series of decomposition-based scene text detection methods has achieved impressive progress by decomposing challenging text regions into pieces and linking them in a bottom-up manner. However, most of them merely focus on linking independent text pieces while the context information is underestimated. In the puzzle game, the solver often put pieces together in a logical way according to the contextual information of each piece, in order to arrive at the correct solution. Inspired by it, we propose a novel decomposition-based method, termed Puzzle Networks (PuzzleNet), to address the challenging scene text detection task in this work. PuzzleNet consists of the Segment Proposal Network (SPN) that predicts the candidate text segments fitting arbitrary shape of text region, and the two-branch Multiple-Similarity Graph Convolutional Network (MSGCN) that models both appearance and geometry correlations between each segment to its contextual ones. By building segments as context graphs, MSGCN effectively employs segment context to predict combinations of segments. Final detections of polygon shape are produced by merging segments according to the predicted combinations. Evaluations on three benchmark datasets, ICDAR15, MSRA-TD500 and SCUT-CTW1500, have demonstrated that our method can achieve better or comparable performance than current state-of-the-arts, which is beneficial from the exploitation of segment context graph.

Deng-Ping Fan, ShengChuan Zhang, Yu-Huan Wu et al.

Human visual system has the strong ability to quick assess the perceptual similarity between two facial sketches. However, existing two widely-used facial sketch metrics, e.g., FSIM and SSIM fail to address this perceptual similarity in this field. Recent study in facial modeling area has verified that the inclusion of both structure and texture has a significant positive benefit for face sketch synthesis (FSS). But which statistics are more important, and are helpful for their success? In this paper, we design a perceptual metric,called Structure Co-Occurrence Texture (Scoot), which simultaneously considers the block-level spatial structure and co-occurrence texture statistics. To test the quality of metrics, we propose three novel meta-measures based on various reliable properties. Extensive experiments demonstrate that our Scoot metric exceeds the performance of prior work. Besides, we built the first large scale (152k judgments) human-perception-based sketch database that can evaluate how well a metric is consistent with human perception. Our results suggest that "spatial structure" and "co-occurrence texture" are two generally applicable perceptual features in face sketch synthesis.

Hao Liu, Jingjing Wu, Jianguo Jiang et al.

Attribute recognition has become crucial because of its wide applications in many computer vision tasks, such as person re-identification. Like many object recognition problems, variations in viewpoints, illumination, and recognition at far distance, all make this task challenging. In this work, we propose a joint CTC-Attention model (JCM), which maps attribute labels into sequences to learn the semantic relationship among attributes. Besides, this network uses neural network to encode images into sequences, and employs connectionist temporal classification (CTC) loss to train the network with the aim of improving the encoding performance of the network. At the same time, it adopts the attention model to decode the sequences, which can realize aligning the sequences and better learning the semantic information from attributes. Extensive experiments on three public datasets, i.e., Market-1501 attribute dataset, Duke attribute dataset and PETA dataset, demonstrate the effectiveness of the proposed method.

Deng-Ping Fan, Cheng Gong, Yang Cao et al.

The existing binary foreground map (FM) measures to address various types of errors in either pixel-wise or structural ways. These measures consider pixel-level match or image-level information independently, while cognitive vision studies have shown that human vision is highly sensitive to both global information and local details in scenes. In this paper, we take a detailed look at current binary FM evaluation measures and propose a novel and effective E-measure (Enhanced-alignment measure). Our measure combines local pixel values with the image-level mean value in one term, jointly capturing image-level statistics and local pixel matching information. We demonstrate the superiority of our measure over the available measures on 4 popular datasets via 5 meta-measures, including ranking models for applications, demoting generic, random Gaussian noise maps, ground-truth switch, as well as human judgments. We find large improvements in almost all the meta-measures. For instance, in terms of application ranking, we observe improvementrangingfrom9.08% to 19.65% compared with other popular measures.

Deng-Ping Fan, ShengChuan Zhang, Yu-Huan Wu et al.

Existing face sketch synthesis (FSS) similarity measures are sensitive to slight image degradation (e.g., noise, blur). However, human perception of the similarity of two sketches will consider both structure and texture as essential factors and is not sensitive to slight ("pixel-level") mismatches. Consequently, the use of existing similarity measures can lead to better algorithms receiving a lower score than worse algorithms. This unreliable evaluation has significantly hindered the development of the FSS field. To solve this problem, we propose a novel and robust style similarity measure called Scoot-measure (Structure CO-Occurrence Texture Measure), which simultaneously evaluates "block-level" spatial structure and co-occurrence texture statistics. In addition, we further propose 4 new meta-measures and create 2 new datasets to perform a comprehensive evaluation of several widely-used FSS measures on two large databases. Experimental results demonstrate that our measure not only provides a reliable evaluation but also achieves significantly improved performance. Specifically, the study indicated a higher degree (78.8%) of correlation between our measure and human judgment than the best prior measure (58.6%). Our code will be made available.