Junfeng Luo

Sixiang Chen, Jianyu Lai, Jialin Gao et al.

Image-to-poster generation is a high-demand task requiring not only local adjustments but also high-level design understanding. Models must generate text, layout, style, and visual elements while preserving semantic fidelity and aesthetic coherence. The process spans two regimes: local editing, where ID-driven generation, rescaling, filling, and extending must preserve concrete visual entities; and global creation, where layout- and style-driven tasks rely on understanding abstract design concepts. These intertwined demands make image-to-poster a multi-dimensional process coupling entity-preserving editing with concept-driven creation under image-prompt control. To address these challenges, we propose PosterOmni, a generalized artistic poster creation framework that unlocks the potential of a base edit model for multi-task image-to-poster generation. PosterOmni integrates the two regimes, namely local editing and global creation, within a single system through an efficient data-distillation-reward pipeline: (i) constructing multi-scenario image-to-poster datasets covering six task types across entity-based and concept-based creation; (ii) distilling knowledge between local and global experts for supervised fine-tuning; and (iii) applying unified PosterOmni Reward Feedback to jointly align visual entity-preserving and aesthetic preference across all tasks. Additionally, we establish PosterOmni-Bench, a unified benchmark for evaluating both local editing and global creation. Extensive experiments show that PosterOmni significantly enhances reference adherence, global composition quality, and aesthetic harmony, outperforming all open-source baselines and even surpassing several proprietary systems.

Junwei Yang, Ke Zhang, Zhaolin Cui et al.

Feature representation via self-supervised learning has reached remarkable success in image-level contrastive learning, which brings impressive performances on image classification tasks. While image-level feature representation mainly focuses on contrastive learning in single instance, it ignores the objective differences between pretext and downstream prediction tasks such as object detection and instance segmentation. In order to fully unleash the power of feature representation on downstream prediction tasks, we propose a new end-to-end self-supervised framework called InsCon, which is devoted to capturing multi-instance information and extracting cell-instance features for object recognition and localization. On the one hand, InsCon builds a targeted learning paradigm that applies multi-instance images as input, aligning the learned feature between corresponding instance views, which makes it more appropriate for multi-instance recognition tasks. On the other hand, InsCon introduces the pull and push of cell-instance, which utilizes cell consistency to enhance fine-grained feature representation for precise boundary localization. As a result, InsCon learns multi-instance consistency on semantic feature representation and cell-instance consistency on spatial feature representation. Experiments demonstrate the method we proposed surpasses MoCo v2 by 1.1% AP^{bb} on COCO object detection and 1.0% AP^{mk} on COCO instance segmentation using Mask R-CNN R50-FPN network structure with 90k iterations, 2.1% APbb on PASCAL VOC objection detection using Faster R-CNN R50-C4 network structure with 24k iterations.

Jinming Su, Wangwang Yang, Junfeng Luo et al.

In order to deal with the task of video panoptic segmentation in the wild, we propose a robust integrated video panoptic segmentation solution. In our solution, we regard the video panoptic segmentation task as a segmentation target querying task, represent both semantic and instance targets as a set of queries, and then combine these queries with video features extracted by neural networks to predict segmentation masks. In order to improve the learning accuracy and convergence speed of the solution, we add additional tasks of video semantic segmentation and video instance segmentation for joint training. In addition, we also add an additional image semantic segmentation model to further improve the performance of semantic classes. In addition, we also add some additional operations to improve the robustness of the model. Extensive experiments on the VIPSeg dataset show that the proposed solution achieves state-of-the-art performance with 50.04\% VPQ on the VIPSeg test set, which is 3rd place on the video panoptic segmentation track of the PVUW Challenge 2023.

Jinming Su, Ruihong Yin, Shuaibin Zhang et al.

In recent years, video semantic segmentation has made great progress with advanced deep neural networks. However, there still exist two main challenges \ie, information inconsistency and computation cost. To deal with the two difficulties, we propose a novel motion-state alignment framework for video semantic segmentation to keep both motion and state consistency. In the framework, we first construct a motion alignment branch armed with an efficient decoupled transformer to capture dynamic semantics, guaranteeing region-level temporal consistency. Then, a state alignment branch composed of a stage transformer is designed to enrich feature spaces for the current frame to extract static semantics and achieve pixel-level state consistency. Next, by a semantic assignment mechanism, the region descriptor of each semantic category is gained from dynamic semantics and linked with pixel descriptors from static semantics. Benefiting from the alignment of these two kinds of effective information, the proposed method picks up dynamic and static semantics in a targeted way, so that video semantic regions are consistently segmented to obtain precise locations with low computational complexity. Extensive experiments on Cityscapes and CamVid datasets show that the proposed approach outperforms state-of-the-art methods and validates the effectiveness of the motion-state alignment framework.

Hongyu Li, Songhao Han, Yue Liao et al.

Understanding real-world videos with complex semantics and long temporal dependencies remains a fundamental challenge in computer vision. Recent progress in multimodal large language models (MLLMs) has demonstrated strong capabilities in vision-language tasks, while reinforcement learning tuning (RLT) has further improved their reasoning abilities. In this work, we explore RLT as a post-training strategy to enhance the video-specific reasoning capabilities of MLLMs. Built upon the Group Relative Policy Optimization (GRPO) framework, we propose a dual-reward formulation that supervises both semantic and temporal reasoning through discrete and continuous reward signals. To facilitate effective preference-based optimization, we introduce a variance-aware data selection strategy based on repeated inference to identify samples that provide informative learning signals. We evaluate our approach across eight representative video understanding tasks, including VideoQA, Temporal Video Grounding, and Grounded VideoQA. Our method consistently outperforms supervised fine-tuning and existing RLT baselines, achieving superior performance with significantly less training data. These results underscore the importance of reward design and data selection in advancing reasoning-centric video understanding with MLLMs. Notably, The initial code release (two months ago) has now been expanded with updates, including optimized reward mechanisms and additional datasets. The latest version is available at https://github.com/appletea233/Temporal-R1 .

SiXiang Chen, Jianyu Lai, Jialin Gao et al.

Generating aesthetic posters is more challenging than simple design images: it requires not only precise text rendering but also the seamless integration of abstract artistic content, striking layouts, and overall stylistic harmony. To address this, we propose PosterCraft, a unified framework that abandons prior modular pipelines and rigid, predefined layouts, allowing the model to freely explore coherent, visually compelling compositions. PosterCraft employs a carefully designed, cascaded workflow to optimize the generation of high-aesthetic posters: (i) large-scale text-rendering optimization on our newly introduced Text-Render-2M dataset; (ii) region-aware supervised fine-tuning on HQ-Poster100K; (iii) aesthetic-text-reinforcement learning via best-of-n preference optimization; and (iv) joint vision-language feedback refinement. Each stage is supported by a fully automated data-construction pipeline tailored to its specific needs, enabling robust training without complex architectural modifications. Evaluated on multiple experiments, PosterCraft significantly outperforms open-source baselines in rendering accuracy, layout coherence, and overall visual appeal-approaching the quality of SOTA commercial systems. Our code, models, and datasets can be found in the Project page: https://ephemeral182.github.io/PosterCraft

Zining Wang, Tongkun Guan, Pei Fu et al.

Multi-modal Large Language Models (MLLMs) have introduced a novel dimension to document understanding, i.e., they endow large language models with visual comprehension capabilities; however, how to design a suitable image-text pre-training task for bridging the visual and language modality in document-level MLLMs remains underexplored. In this study, we introduce a novel visual-language alignment method that casts the key issue as a Visual Question Answering with Mask generation (VQAMask) task, optimizing two tasks simultaneously: VQA-based text parsing and mask generation. The former allows the model to implicitly align images and text at the semantic level. The latter introduces an additional mask generator (discarded during inference) to explicitly ensure alignment between visual texts within images and their corresponding image regions at a spatially-aware level. Together, they can prevent model hallucinations when parsing visual text and effectively promote spatially-aware feature representation learning. To support the proposed VQAMask task, we construct a comprehensive image-mask generation pipeline and provide a large-scale dataset with 6M data (MTMask6M). Subsequently, we demonstrate that introducing the proposed mask generation task yields competitive document-level understanding performance. Leveraging the proposed VQAMask, we introduce Marten, a training-efficient MLLM tailored for document-level understanding. Extensive experiments show that our Marten consistently achieves significant improvements among 8B-MLLMs in document-centric tasks. Code and datasets are available at https://github.com/PriNing/Marten.

Tongkun Guan, Zining Wang, Pei Fu et al.

In recent years, general visual foundation models (VFMs) have witnessed increasing adoption, particularly as image encoders for popular multi-modal large language models (MLLMs). However, without semantically fine-grained supervision, these models still encounter fundamental prediction errors in the context of downstream text-image-related tasks, i.e., perception, understanding and reasoning with images containing small and dense texts. To bridge this gap, we develop TokenOCR, the first token-level visual foundation model specifically tailored for text-image-related tasks, designed to support a variety of traditional downstream applications. To facilitate the pretraining of TokenOCR, we also devise a high-quality data production pipeline that constructs the first token-level image text dataset, TokenIT, comprising 20 million images and 1.8 billion token-mask pairs. Furthermore, leveraging this foundation with exceptional image-as-text capability, we seamlessly replace previous VFMs with TokenOCR to construct a document-level MLLM, TokenVL, for VQA-based document understanding tasks. Finally, extensive experiments demonstrate the effectiveness of TokenOCR and TokenVL. Code, datasets, and weights will be available at https://github.com/Token-family/TokenFD.

Jinming Su, Ruihong Yin, Xingyue Chen et al.

Recently, instance segmentation has made great progress with the rapid development of deep neural networks. However, there still exist two main challenges including discovering indistinguishable objects and modeling the relationship between instances. To deal with these difficulties, we propose a novel object mining framework for instance segmentation. In this framework, we first introduce the semantics perceiving subnetwork to capture pixels that may belong to an obvious instance from the bottom up. Then, we propose an object excavating mechanism to discover indistinguishable objects. In the mechanism, preliminary perceived semantics are regarded as original instances with classifications and locations, and then indistinguishable objects around these original instances are mined, which ensures that hard objects are fully excavated. Next, an instance purifying strategy is put forward to model the relationship between instances, which pulls the similar instances close and pushes away different instances to keep intra-instance similarity and inter-instance discrimination. In this manner, the same objects are combined as the one instance and different objects are distinguished as independent instances. Extensive experiments on the COCO dataset show that the proposed approach outperforms state-of-the-art methods, which validates the effectiveness of the proposed object mining framework.

Wangwang Yang, Jinming Su, Yiting Duan et al.

Video object segmentation (VOS) has made significant progress with the rise of deep learning. However, there still exist some thorny problems, for example, similar objects are easily confused and tiny objects are difficult to be found. To solve these problems and further improve the performance of VOS, we propose a simple yet effective solution for this task. In the solution, we first analyze the distribution of the Youtube-VOS dataset and supplement the dataset by introducing public static and video segmentation datasets. Then, we improve three network architectures with different characteristics and train several networks to learn the different characteristics of objects in videos. After that, we use a simple way to integrate all results to ensure that different models complement each other. Finally, subtle post-processing is carried out to ensure accurate video object segmentation with precise boundaries. Extensive experiments on Youtube-VOS dataset show that the proposed solution achieves the state-of-the-art performance with an 86.1% overall score on the YouTube-VOS 2022 test set, which is 5th place on the video object segmentation track of the Youtube-VOS Challenge 2022.

Yifu Chen, Bingchen Huang, Zhiling Wang et al.

In-context learning (ICL) has become a classic approach for enabling LLMs to handle various tasks based on a few input-output examples. The effectiveness of ICL heavily relies on the quality of these examples, and previous works which focused on enhancing example retrieval capabilities have achieved impressive performances. However, two challenges remain in retrieving high-quality examples: (1) Difficulty in distinguishing cross-task data distributions, (2) Difficulty in making the fine-grained connection between retriever output and feedback from LLMs. In this paper, we propose a novel framework called TDR. TDR decouples the ICL examples from different tasks, which enables the retrieval module to retrieve examples specific to the target task within a multi-task dataset. Furthermore, TDR models fine-grained feedback from LLMs to supervise and guide the training of the retrieval module, which helps to retrieve high-quality examples. We conducted extensive experiments on a suite of 30 NLP tasks, the results demonstrate that TDR consistently improved results across all datasets and achieves state-of-the-art performance. Meanwhile, our approach is a plug-and-play method, which can be easily combined with various LLMs to improve example retrieval abilities for ICL. The code is available at https://github.com/Nnn-s/TDR.

Jinming Su, Songen Gu, Yiting Duan et al.

Text-to-image generation has made remarkable progress with the emergence of diffusion models. However, it is still a difficult task to generate images for street views based on text, mainly because the road topology of street scenes is complex, the traffic status is diverse and the weather condition is various, which makes conventional text-to-image models difficult to deal with. To address these challenges, we propose a novel controllable text-to-image framework, named \textbf{Text2Street}. In the framework, we first introduce the lane-aware road topology generator, which achieves text-to-map generation with the accurate road structure and lane lines armed with the counting adapter, realizing the controllable road topology generation. Then, the position-based object layout generator is proposed to obtain text-to-layout generation through an object-level bounding box diffusion strategy, realizing the controllable traffic object layout generation. Finally, the multiple control image generator is designed to integrate the road topology, object layout and weather description to realize controllable street-view image generation. Extensive experiments show that the proposed approach achieves controllable street-view text-to-image generation and validates the effectiveness of the Text2Street framework for street views.

Hengyu Shi, Junhao Su, Junfeng Luo et al.

Conditional layout generation aims to automatically generate visually appealing and semantically coherent layouts from user-defined constraints. While recent methods based on generative models have shown promising results, they typically require substantial amounts of training data or extensive fine-tuning, limiting their versatility and practical applicability. Alternatively, some training-free approaches leveraging in-context learning with Large Language Models (LLMs) have emerged, but they often suffer from limited reasoning capabilities and overly simplistic ranking mechanisms, which restrict their ability to generate consistently high-quality layouts. To this end, we propose LayoutCoT, a novel approach that leverages the reasoning capabilities of LLMs through a combination of Retrieval-Augmented Generation (RAG) and Chain-of-Thought (CoT) techniques. Specifically, LayoutCoT transforms layout representations into a standardized serialized format suitable for processing by LLMs. A Layout-aware RAG is used to facilitate effective retrieval and generate a coarse layout by LLMs. This preliminary layout, together with the selected exemplars, is then fed into a specially designed CoT reasoning module for iterative refinement, significantly enhancing both semantic coherence and visual quality. We conduct extensive experiments on five public datasets spanning three conditional layout generation tasks. Experimental results demonstrate that LayoutCoT achieves state-of-the-art performance without requiring training or fine-tuning. Notably, our CoT reasoning module enables standard LLMs, even those without explicit deep reasoning abilities, to outperform specialized deep-reasoning models such as deepseek-R1, highlighting the potential of our approach in unleashing the deep reasoning capabilities of LLMs for layout generation tasks.

Chen Duan, Qianyi Jiang, Pei Fu et al.

In the field of scene text spotting, previous OCR methods primarily relied on image encoders and pre-trained text information, but they often overlooked the advantages of incorporating human language instructions. To address this gap, we propose InstructOCR, an innovative instruction-based scene text spotting model that leverages human language instructions to enhance the understanding of text within images. Our framework employs both text and image encoders during training and inference, along with instructions meticulously designed based on text attributes. This approach enables the model to interpret text more accurately and flexibly. Extensive experiments demonstrate the effectiveness of our model and we achieve state-of-the-art results on widely used benchmarks. Furthermore, the proposed framework can be seamlessly applied to scene text VQA tasks. By leveraging instruction strategies during pre-training, the performance on downstream VQA tasks can be significantly improved, with a 2.6% increase on the TextVQA dataset and a 2.1% increase on the ST-VQA dataset. These experimental results provide insights into the benefits of incorporating human language instructions for OCR-related tasks.

Junhao Su, Yuanliang Wan, Junwei Yang et al.

Tool-augmented large language models (LLMs) are usually trained with supervised imitation or coarse-grained reinforcement learning that optimizes single tool calls. Current self-reflection practices rely on heuristic prompts or one-way reasoning: the model is urged to 'think more' instead of learning error diagnosis and repair. This is fragile in multi-turn interactions; after a failure the model often repeats the same mistake. We propose structured reflection, which turns the path from error to repair into an explicit, controllable, and trainable action. The agent produces a short yet precise reflection: it diagnoses the failure using evidence from the previous step and then proposes a correct, executable follow-up call. For training we combine DAPO and GSPO objectives with a reward scheme tailored to tool use, optimizing the stepwise strategy Reflect, then Call, then Final. To evaluate, we introduce Tool-Reflection-Bench, a lightweight benchmark that programmatically checks structural validity, executability, parameter correctness, and result consistency. Tasks are built as mini trajectories of erroneous call, reflection, and corrected call, with disjoint train and test splits. Experiments on BFCL v3 and Tool-Reflection-Bench show large gains in multi-turn tool-call success and error recovery, and a reduction of redundant calls. These results indicate that making reflection explicit and optimizing it directly improves the reliability of tool interaction and offers a reproducible path for agents to learn from failure.

Junhao Su, Feiyu Zhu, Hengyu Shi et al.

Deep learning typically relies on end-to-end backpropagation for training, a method that inherently suffers from issues such as update locking during parameter optimization, high GPU memory consumption, and a lack of biological plausibility. In contrast, supervised local learning seeks to mitigate these challenges by partitioning the network into multiple local blocks and designing independent auxiliary networks to update each block separately. However, because gradients are propagated solely within individual local blocks, performance degradation occurs, preventing supervised local learning from supplanting end-to-end backpropagation. To address these limitations and facilitate inter-block information flow, we propose the Momentum Auxiliary Network++ (MAN++). MAN++ introduces a dynamic interaction mechanism by employing the Exponential Moving Average (EMA) of parameters from adjacent blocks to enhance communication across the network. The auxiliary network, updated via EMA, effectively bridges the information gap between blocks. Notably, we observed that directly applying EMA parameters can be suboptimal due to feature discrepancies between local blocks. To resolve this issue, we introduce a learnable scaling bias that balances feature differences, thereby further improving performance. We validate MAN++ through extensive experiments on tasks that include image classification, object detection, and image segmentation, utilizing multiple network architectures. The experimental results demonstrate that MAN++ achieves performance comparable to end-to-end training while significantly reducing GPU memory usage. Consequently, MAN++ offers a novel perspective for supervised local learning and presents a viable alternative to conventional training methods.

Junjie Hu, Tianyang Han, Kai Ma et al.

Recent subject-driven image customization has achieved significant advancements in fidelity, yet fine-grained instance-level spatial control remains elusive, hindering broader real-world application. This limitation is mainly attributed to the absence of scalable datasets that bind identity with precise positional cues. To this end, we introduce PositionIC, a unified framework that enforces position and identity consistency for multi-subject customization. We construct a scalable synthesis pipeline that employs a bidirectional generation paradigm to eliminate subject drift and maintain semantic coherence. On top of these data, we design a lightweight positional modulation operation that decouples spatial embeddings among subjects, enabling independent, accurate placement while preserving visual fidelity. Extensive experiments demonstrate that our approach can achieve precise spatial control while maintaining high consistency in image customization tasks. PositionIC paves the way for controllable, high-fidelity image customization in open-world, multi-entity scenarios and will be released to foster further research.

Pei Fu, Tongkun Guan, Zining Wang et al.

The recent emergence of Multi-modal Large Language Models (MLLMs) has introduced a new dimension to the Text-rich Image Understanding (TIU) field, with models demonstrating impressive and inspiring performance. However, their rapid evolution and widespread adoption have made it increasingly challenging to keep up with the latest advancements. To address this, we present a systematic and comprehensive survey to facilitate further research on TIU MLLMs. Initially, we outline the timeline, architecture, and pipeline of nearly all TIU MLLMs. Then, we review the performance of selected models on mainstream benchmarks. Finally, we explore promising directions, challenges, and limitations within the field.

Jianyu Lai, Sixiang Chen, Jialin Gao et al.

Recent advancements in the text-rendering capabilities of image generation models have made the end-to-end creation of graphic design content, such as posters, increasingly feasible. However, existing reward models fall short of accurately assessing design quality, as they primarily focus on global image aesthetics while overlooking the critical dimensions of typography and layout. Furthermore, the scarcity of domain-specific preference data remains a significant bottleneck, which limits the further development of graphic design evaluation and generation. To bridge this gap, we introduce an automated pipeline to construct a high-quality dataset of 70k poster preferences by leveraging the consensus of multiple Multi-modal Large Language Models (MLLMs) to simulate human-like judgment. Utilizing this dataset, we develop PosterReward, a reward model specifically designed for high-precision poster assessment through a cascaded, multi-stage training strategy. We also provide multiple variants of the model to cater to different application scenarios. Finally, we introduce PosterRewardBench and PosterBench to evaluate the performance of existing reward models in poster assessment and the generation capabilities of current text-to-image models in poster creation, respectively.

Tianyang Han, Junhao Su, Junjie Hu et al.

Text-to-image (T2I) models today are capable of producing photorealistic, instruction-following images, yet they still frequently fail on prompts that require implicit world knowledge. Existing evaluation protocols either emphasize compositional alignment or rely on single-round VQA-based scoring, leaving critical dimensions such as knowledge grounding, multi-physics interactions, and auditable evidence-substantially undertested. To address these limitations, we introduce PicWorld, the first comprehensive benchmark that assesses the grasp of implicit world knowledge and physical causal reasoning of T2I models. This benchmark consists of 1,100 prompts across three core categories. To facilitate fine-grained evaluation, we propose PW-Agent, an evidence-grounded multi-agent evaluator to hierarchically assess images on their physical realism and logical consistency by decomposing prompts into verifiable visual evidence. We conduct a thorough analysis of 17 mainstream T2I models on PicWorld, illustrating that they universally exhibit a fundamental limitation in their capacity for implicit world knowledge and physical causal reasoning to varying degrees. The findings highlight the need for reasoning-aware, knowledge-integrative architectures in future T2I systems.

Jiawei Zhao, Qixing Jiang, Xuede Li et al.

Birds-Eye-View (BEV) segmentation aims to establish a spatial mapping from the perspective view to the top view and estimate the semantic maps from monocular images. Recent studies have encountered difficulties in view transformation due to the disruption of BEV-agnostic features in image space. To tackle this issue, we propose a novel FocusBEV framework consisting of $(i)$ a self-calibrated cross view transformation module to suppress the BEV-agnostic image areas and focus on the BEV-relevant areas in the view transformation stage, $(ii)$ a plug-and-play ego-motion-based temporal fusion module to exploit the spatiotemporal structure consistency in BEV space with a memory bank, and $(iii)$ an occupancy-agnostic IoU loss to mitigate both semantic and positional uncertainties. Experimental evidence demonstrates that our approach achieves new state-of-the-art on two popular benchmarks,\ie, 29.2\% mIoU on nuScenes and 35.2\% mIoU on Argoverse.

Feiyu Zhu, Yuming Zhang, Xiuyuan Guo et al.

Local learning offers an alternative to traditional end-to-end back-propagation in deep neural networks, significantly reducing GPU memory consumption. Although it has shown promise in image classification tasks, its extension to other visual tasks has been limited. This limitation arises primarily from two factors: 1) architectures designed specifically for classification are not readily adaptable to other tasks, which prevents the effective reuse of task-specific knowledge from architectures tailored to different problems; 2) these classification-focused architectures typically lack cross-scale feature communication, leading to degraded performance in tasks like object detection and super-resolution. To address these challenges, we propose the Feature Bank Augmented auxiliary network (FBA), which introduces a simplified design principle and incorporates a feature bank to enhance cross-task adaptability and communication. This work represents the first successful application of local learning methods beyond classification, demonstrating that FBA not only conserves GPU memory but also achieves performance on par with end-to-end approaches across multiple datasets for various visual tasks.

Jinming Su, Chao Chen, Ke Zhang et al.

Recently, lane detection has made great progress with the rapid development of deep neural networks and autonomous driving. However, there exist three mainly problems including characterizing lanes, modeling the structural relationship between scenes and lanes, and supporting more attributes (e.g., instance and type) of lanes. In this paper, we propose a novel structure guided framework to solve these problems simultaneously. In the framework, we first introduce a new lane representation to characterize each instance. Then a topdown vanishing point guided anchoring mechanism is proposed to produce intensive anchors, which efficiently capture various lanes. Next, multi-level structural constraints are used to improve the perception of lanes. In the process, pixel-level perception with binary segmentation is introduced to promote features around anchors and restore lane details from bottom up, a lane-level relation is put forward to model structures (i.e., parallel) around lanes, and an image-level attention is used to adaptively attend different regions of the image from the perspective of scenes. With the help of structural guidance, anchors are effectively classified and regressed to obtain precise locations and shapes. Extensive experiments on public benchmark datasets show that the proposed approach outperforms state-of-the-art methods with 117 FPS on a single GPU.

Mingyuan Fan, Shenqi Lai, Junshi Huang et al.

BiSeNet has been proved to be a popular two-stream network for real-time segmentation. However, its principle of adding an extra path to encode spatial information is time-consuming, and the backbones borrowed from pretrained tasks, e.g., image classification, may be inefficient for image segmentation due to the deficiency of task-specific design. To handle these problems, we propose a novel and efficient structure named Short-Term Dense Concatenate network (STDC network) by removing structure redundancy. Specifically, we gradually reduce the dimension of feature maps and use the aggregation of them for image representation, which forms the basic module of STDC network. In the decoder, we propose a Detail Aggregation module by integrating the learning of spatial information into low-level layers in single-stream manner. Finally, the low-level features and deep features are fused to predict the final segmentation results. Extensive experiments on Cityscapes and CamVid dataset demonstrate the effectiveness of our method by achieving promising trade-off between segmentation accuracy and inference speed. On Cityscapes, we achieve 71.9% mIoU on the test set with a speed of 250.4 FPS on NVIDIA GTX 1080Ti, which is 45.2% faster than the latest methods, and achieve 76.8% mIoU with 97.0 FPS while inferring on higher resolution images.