Vision-Centric Representation-Efficient Fine-Tuning for Robust Universal Foreground Segmentation

Foreground segmentation is crucial for scene understanding, yet parameter-efficient fine-tuning (PEFT) of vision foundation models (VFMs) often fails in complex scenarios, such as camouflage and infrared imagery. We attribute this challenge to the inherent texture bias in VFMs, which is exacerbated during fine-tuning and limits generalization in texture-sparse environments. To address this, we propose Ladder Shape-bias Representation Side-tuning (LSR-ST), a lightweight PEFT framework that enhances model robustness by introducing shape-biased inductive priors. LSR-ST captures shape-aware features using a simple HDConv Block, which integrates large-kernel attention and residual learning. The method satisfies three key conditions for inducing shape bias: large receptive fields, multi-order feature interactions, and sparse connectivity. Our analysis reveals that these improvements stem from representation efficiency-the ability to extract task-relevant, structurally grounded features while minimizing redundancy. We formalize this concept via Information Bottleneck theory and advocate for it as a key PEFT objective. Unlike traditional NLP paradigms that focus on optimizing parameters and memory, visual tasks require models that extract task-defined semantics, rather than just relying on pre-encoded features. This shift enables our approach to move beyond conventional trade-offs, offering more robust and generalizable solutions for vision tasks. With minimal changes to SAM2-UNet, LSR-ST achieves consistent improvements across 17 datasets and 6 tasks using only 4.719M trainable parameters. These results highlight the potential of representation efficiency for robust and adaptable VFMs within complex visual environments.

Optimizing Local-Global Dependencies for Accurate 3D Human Pose Estimation

Transformer-based methods have recently achieved significant success in 3D human pose estimation, owing to their strong ability to model long-range dependencies. However, relying solely on the global attention mechanism is insufficient for capturing the fine-grained local details, which are crucial for accurate pose estimation. To address this, we propose SSR-STF, a dual-stream model that effectively integrates local features with global dependencies to enhance 3D human pose estimation. Specifically, we introduce SSRFormer, a simple yet effective module that employs the skeleton selective refine attention (SSRA) mechanism to capture fine-grained local dependencies in human pose sequences, complementing the global dependencies modeled by the Transformer. By adaptively fusing these two feature streams, SSR-STF can better learn the underlying structure of human poses, overcoming the limitations of traditional methods in local feature extraction. Extensive experiments on the Human3.6M and MPI-INF-3DHP datasets demonstrate that SSR-STF achieves state-of-the-art performance, with P1 errors of 37.4 mm and 13.2 mm respectively, outperforming existing methods in both accuracy and generalization. Furthermore, the motion representations learned by our model prove effective in downstream tasks such as human mesh recovery. Codes are available at https://github.com/poker-xu/SSR-STF.

Learning Dynamic Local Context Representations for Infrared Small Target Detection

Infrared small target detection (ISTD) is challenging due to complex backgrounds, low signal-to-clutter ratios, and varying target sizes and shapes. Effective detection relies on capturing local contextual information at the appropriate scale. However, small-kernel CNNs have limited receptive fields, leading to false alarms, while transformer models, with global receptive fields, often treat small targets as noise, resulting in miss-detections. Hybrid models struggle to bridge the semantic gap between CNNs and transformers, causing high complexity.To address these challenges, we propose LCRNet, a novel method that learns dynamic local context representations for ISTD. The model consists of three components: (1) C2FBlock, inspired by PDE solvers, for efficient small target information capture; (2) DLC-Attention, a large-kernel attention mechanism that dynamically builds context and reduces feature redundancy; and (3) HLKConv, a hierarchical convolution operator based on large-kernel decomposition that preserves sparsity and mitigates the drawbacks of dilated convolutions. Despite its simplicity, with only 1.65M parameters, LCRNet achieves state-of-the-art (SOTA) performance.Experiments on multiple datasets, comparing LCRNet with 33 SOTA methods, demonstrate its superior performance and efficiency.