Depth-wise pruning accelerates LLM inference in resource-constrained scenarios but suffers from performance degradation due to indiscriminate removal of entire Transformer layers. This paper reveals ``Patch-Like'' redundancy across layers via correlation analysis of the outputs of different layers in reproducing kernel Hilbert space, demonstrating consecutive layers exhibit high functional similarity. Building on this observation, this paper proposes Sliding-Window Merging (SWM) - a dynamic compression method that selects consecutive layers from top to bottom using a pre-defined similarity threshold, and compacts patch-redundant layers through a parameter consolidation, thereby simplifying the model structure while maintaining its performance. Extensive experiments on LLMs with various architectures and different parameter scales show that our method outperforms existing pruning techniques in both zero-shot inference performance and retraining recovery quality after pruning. In particular, in the experiment with 35\% pruning on the Vicuna-7B model, our method achieved a 1.654\% improvement in average performance on zero-shot tasks compared to the existing method. Moreover, we further reveal the potential of combining depth pruning with width pruning to enhance the pruning effect.

Precise vessel segmentation is vital for clinical applications such as diagnosis and surgical planning but remains challenging due to thin, branching geometries and low texture contrast. Although foundation models such as the Segment Anything Model (SAM) show strong performance in general segmentation tasks, they remain suboptimal for vascular structures. In this work, we present VesSAM, a powerful and efficient framework tailored for 2D vessel segmentation. VesSAM integrates three core modules: a convolutional adapter that enhances local texture features, a multi-prompt encoder that fuses anatomical cues via hierarchical cross-attention, and a lightweight mask decoder that reduces jagged artifacts. We also introduce an automated pipeline to generate structured multi-prompt annotations, and curate a diverse benchmark dataset spanning 8 datasets across 5 imaging modalities. Extensive experiments show that VesSAM surpasses state-of-the-art PEFT-based SAM variants by over$\text{1 0 \%}$Dice and 13% IoU, while maintaining competitive accuracy to fully fine-tuned methods with far fewer parameters. VesSAM also generalizes well to out-of-distribution (OoD) settings, outperforming all baselines in average OoD Dice and IoU.

Laser speckle contrast imaging (LSCI) is widely used for in vivo real-time detection and analysis of local blood flow microcirculation due to its non-invasive ability and excellent spatial and temporal resolution. However, vascular segmentation of LSCI images still faces a lot of difficulties due to numerous specific noises caused by the complexity of blood microcirculation's structure and irregular vascular aberrations in diseased regions. In addition, the difficulties of LSCI image data annotation have hindered the application of deep learning methods based on supervised learning in the field of LSCI image vascular segmentation. To tackle these difficulties, we propose a robust weakly supervised learning method, which selects the threshold combinations and processing flows instead of labor-intensive annotation work to construct the ground truth of the dataset, and design a deep neural network, FURNet, based on UNet++ and ResNeXt. The model obtained from training achieves high-quality vascular segmentation and captures multi-scene vascular features on both constructed and unknown datasets with good generalization. Furthermore, we intravital verified the availability of this method on a tumor before and after embolization treatment. This work provides a new approach for realizing LSCI vascular segmentation and also makes a new application-level advance in the field of artificial intelligence-assisted disease diagnosis.