The paper introduces the Consistency Memory Bank (COMB), a label‑free virtual staining framework designed to process gigapixel Whole Slide Images without the memory bottlenecks of patch‑based deep learning. COMB decouples context storage from computation, using a dynamic retrieval mechanism to fetch feature representations from adjacent tiles, local padding to resolve spatial discontinuities, and neighbor‑aware channel attention to stabilize statistical drift. The method achieves superior perceptual fidelity and tiling consistency compared to state‑of‑the‑art baselines, and its improved continuity suggests downstream benefits for tumor segmentation.
arXiv:2609.00396v1 Announce Type: new Abstract: Histopathological whole slide images (WSIs) are central to cancer diagnosis, but their gigapixel scale, tissue heterogeneity, weak slide-level supervis...
MoSSGate is a plug‑and‑play module for U‑Net that improves skin lesion segmentation by combining boundary‑aware spatial gating, an external memory modulator, and parallel 2D state‑space modeling for efficient global context aggregation. The design limits long‑range propagation to informative regions, adapts dynamically to each sample, and preserves sharp lesion boundaries while keeping computational cost low. Experiments on ISIC 2017 and 2018 show state‑of‑the‑art accuracy (86.3%/85.9% mIoU, 92.6%/90.6% Dice) with fewer FLOPs than most CNN baselines.
SWITi is a test-time method for reducing artifacts in tiled predictions, particularly for neural networks that learn posterior distributions from which solutions are sampled at inference time. Tiled predictions are unavoidable for large image data, and artifacts arise whenever tiles are smaller than a network's receptive field and when tiles are independent posterior samples.
arXiv:2607. 18990v1 Announce Type: cross Abstract: SWITi is a test-time method for reducing artifacts in tiled predictions, particularly for neural networks that learn posterior distributions from which solutions are sampled at inference time.
arXiv:2511. 01143v2 Announce Type: replace-cross Abstract: Early and accurate segmentation of colorectal polyps is critical for reducing colorectal cancer mortality, which has been extensively explored by academia and industry.
arXiv:2606. 06864v1 Announce Type: cross Abstract: Multiple instance learning (MIL) has become a standard paradigm for whole slide image (WSI) analysis in digital pathology, as it enables slide-level prediction without dense annotations.
arXiv:2608. 15537v1 Announce Type: cross Abstract: Accurate boundary delineation remains a persistent challenge in dermoscopic image segmentation because of blurred lesion margins, heterogeneous textures, and complex background artifacts.
AtlasPatch is a scalable, high‑throughput whole‑slide image preprocessing method that uses a foundation‑model‑based tissue detector operating at thumbnail resolution. By updating only 0.076% of the SAM2 model weights and leveraging a curated dataset of 30,000 thumbnail‑mask pairs, it generates accurate tissue masks and directly produces patch coordinates at the desired magnification, eliminating repeated patch‑level inference. The approach achieves 0.986 precision, is up to 16× faster than existing deep‑learning methods, and maintains downstream multiple‑instance learning performance across six slide‑level classification tasks.
Whole-slide image (WSI) analysis remains computationally challenging due to the extremely large spatial resolution of slides and the sparse distribution of tumour regions. We propose an end-to-end reinforcement learning framework for sequential tumour segmentation directly on WSIs.
Medical image segmentation relies on the ability of encoder-decoder architectures to translate rich feature representations into accurate pixel-level predictions under challenging conditions such as low contrast, structural ambiguity, and scale variability. While recent advances in large-scale pretraining and transformer-based encoders have substantially improved feature extraction, segmentation accuracy remains constrained by decoder design, particularly in terms of cross-scale alignment, contextual integration, and boundary preservation.
StainPresetNet is a new framework for stain normalization that combines structural preservation with dataset-level color mapping while remaining computationally efficient. It uses pixel-wise normalization guided by preset reference images, allowing multi-directional adaptability without retraining. Experiments on cytopathology and histopathology datasets show that it outperforms conventional methods in color mapping accuracy, improves classifier generalization, and cuts computational overhead by 90% compared to existing deep learning approaches.
arXiv:2607. 18218v1 Announce Type: cross Abstract: Foundation models have emerged as a driving force in computational pathology, with the potential to transform cancer diagnosis, prognosis, and treatment selection by learning transferable representations from large-scale histopathology data.