In recent years, point regression-based crowd counting has demonstrated remarkable advantages for accurate counting in highly congested scenes. However, existing single-pass regression frameworks lack an explicit self-correction mechanism, making it difficult to rectify localization shifts or even false positives induced by local visual ambiguity in heavily occluded regions. Moreover, straightforward regression predictions struggle to capture complex cluster-level higher-order relations in dense crowds, limiting the model’s ability to leverage global context to resolve semantic confusion. To address these issues, we reformulate point regression as an iterative optimization process and propose a Hypergraph-induced Iterative Reasoning Network, HIRNet. Specifically, we first design a 2D bidirectional hypergraph iterative reasoning module (HIR-2D), which treats intermediate-round regression points as an explicit structural prior and dynamically constructs bidirectional hypergraphs along the spatial and channel dimensions. By performing higher-order feature aggregation in both dimensions, HIR-2D enforces long-range spatial consistency and recalibrates channel features, thereby effectively suppressing background distractions. Second, to mitigate the matching jitter of regressed point sets across iterations, we introduce an Anchor-Matching Iterative Deep Supervision strategy that locks the bipartite matching from the final round to ensure consistent gradient directions across rounds. Furthermore, we devise a Monotonic Improvement Loss that explicitly constrains the model to progressively reduce geometric errors throughout iterations, stabilizing the reasoning dynamics. Extensive experiments on three mainstream datasets demonstrate that our method substantially outperforms current advanced approaches.

Medical image segmentation plays an important role in clinical decision making, treatment planning, and disease tracking. However, it still faces two major challenges. On the one hand, there is often a "soft boundary" between foreground and background in medical images, with poor illumination and low contrast further reducing the distinguishability of foreground and background within the image. On the other hand, co-occurrence phenomena are widespread in medical images, and learning these features is misleading to the model's judgment. To address these challenges, we propose a general framework called Contrast-Driven Medical Image Segmentation (ConDSeg). First, we develop a contrastive training strategy called Consistency Reinforcement. It is designed to improve the encoder's robustness in various illumination and contrast scenarios, enabling the model to extract high-quality features even in adverse environments. Second, we introduce a Semantic Information Decoupling module, which is able to decouple features from the encoder into foreground, background, and uncertainty regions, gradually acquiring the ability to reduce uncertainty during training. The Contrast-Driven Feature Aggregation module then contrasts the foreground and background features to guide multi-level feature fusion and key feature enhancement, further distinguishing the entities to be segmented. We also propose a Size-Aware Decoder to solve the scale singularity of the decoder. It accurately locate entities of different sizes in the image, thus avoiding erroneous learning of co-occurrence features. Extensive experiments on five datasets across three scenarios demonstrate the state-of-the-art performance of our method, proving its advanced nature and general applicability to various medical image segmentation scenarios.

Currently many benchmarks have been proposed to evaluate the perception ability of the Large Vision-Language Models (LVLMs). However, most benchmarks conduct questions by selecting images from existing datasets, resulting in the potential data leakage. Besides, these benchmarks merely focus on evaluating LVLMs on the realistic style images and clean scenarios, leaving the multi-stylized images and noisy scenarios unexplored. In response to these challenges, we propose a dynamic and scalable benchmark named Dysca for evaluating LVLMs by leveraging synthesis images. Specifically, we leverage Stable Diffusion and design a rule-based method to dynamically generate novel images, questions and the corresponding answers. We consider 51 kinds of image styles and evaluate the perception capability in 20 subtasks. Moreover, we conduct evaluations under 4 scenarios (i.e., Clean, Corruption, Print Attacking and Adversarial Attacking) and 3 question types (i.e., Multi-choices, True-or-false and Free-form). Thanks to the generative paradigm, Dysca serves as a scalable benchmark for easily adding new subtasks and scenarios. A total of 24 advanced open-source LVLMs and 2 close-source LVLMs are evaluated on Dysca, revealing the drawbacks of current LVLMs. The benchmark is released in anonymous github page \url{https://github.com/Benchmark-Dysca/Dysca}.

Small object detection in remote-sensing images is essential, yet challenging due to the unique characteristics of small objects. On the one hand, it can be difficult to distinguish small objects from complex backgrounds. On the other hand, due to their minuscule size, small objects are also easily submerged during feature fusion. This letter proposes a novel detection method called sparse outlook network (SOLO-Net) to address these issues. First, we propose a top-k sparse outlook (TKSO) attention module and the sparse outlook path aggregation network (SOLO-PAN) as the fundamental component of SOLO-Net. This module improves the performance of the path aggregation network (PANet), thereby enhancing the ability of the model to focus on small objects. Second, we propose a Sigmoid-intersection over union (IoU) loss function specifically designed for small objects, accelerating model convergence and improving detection performance. Finally, we evaluate our model on the RSOD and DIOR datasets, achieving mean average precision (mAP) scores of 93.8% and 74.5%, respectively.

In recent years, crowd counting has garnered increasing attention due to its wide range of societal applications. However, the vast differences in crowd density distributions across various areas within a scene make it difficult for existing crowd counting methods to achieve satisfactory performance. To address this issue, this letter introduces a novel method called Dynamic Density-Region-Aware Network (DDRANet). First, a Density-Aware Module (DAM) is designed to detect density variations and dynamically segment the image into multiple density regions. Subsequently, the Region Information Encoder (RIE) transforms the obtained density region map into a latent space to enhance the information of density regions. Then, through a specially designed Region Attention Block (RAB), the encoded density region map guides the model to focus on different density regions separately, enabling the model to adapt to the density variations of the crowd within the image. Furthermore, we combine Gaussian kernel function and Density Peak Clustering to automatically generate ground truth of density region maps, supervising the optimization of DAM. Extensive experimental results on three datasets demonstrate DDRANet's superior performance over other state-of-the-art methods.