Abstract:This work proposes a novel approach beyond supervised learning for effective pathological image analysis, addressing the challenge of limited robust labeled data. Pathological diagnosis of diseases like cancer has conventionally relied on the evaluation of morphological features by physicians and pathologists. However, recent advancements in compute-aided diagnosis (CAD) systems are gaining significant attention as diagnostic support tools. Although the advancement of deep learning has improved CAD significantly, segmentation models typically require large pixel-level annotated dataset, and such labeling is expensive. Existing studies not based on supervised approaches still struggle with limited generalization, and no practical approach has emerged yet. To address this issue, we present a weakly supervised semantic segmentation (WSSS) model by combining class activation map and Segment Anything Model (SAM)-based pseudo-labeling. For effective pretraining, we adopt the SAM-a foundation model that is pretrained on large datasets and operates in zero-shot configurations using only coarse prompts. The proposed approach transfer enhanced Attention Dropout Layer's knowledge to SAM, thereby generating pseudo-labels. To demonstrate the superiority of the proposed method, experimental studies are conducted on histopathological breast cancer datasets. The proposed method outperformed other WSSS methods across three datasets, demonstrating its efficiency by achieving this with only 12GB of GPU memory during training. Our code is available at : <a class="link-external link-https" href="https://github.com/QI-NemoSong/EP-SAM" rel="external noopener nofollow">this https URL</a>

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the problem of **insufficient labeled data** in pathological image segmentation, especially in the pathological diagnosis of diseases such as cancer. Specifically: 1. **Scarcity of labeled data**: - The segmentation task of pathological images usually requires a large amount of pixel - level labeled data, but the acquisition of these data is very expensive and time - consuming, especially in the medical field, because it requires professional pathologists to label. 2. **Limitations of weakly - supervised learning**: - Although the existing weakly - supervised semantic segmentation (WSSS) methods reduce the dependence on a large amount of labeled data, they still have limitations in generalization ability and accuracy, especially when dealing with pathological images with fuzzy boundaries. 3. **Limitations of Segment Anything Model (SAM)**: - SAM performs excellently on natural images, but on medical images, due to the indistinct distinction between foreground and background, its zero - shot performance is poor, and the selection of different prompts will lead to significant performance differences. ### The method proposed in the paper To solve the above problems, the paper proposes a new method named **EP - SAM**, which combines the class activation map (CAM) and the pseudo - label generation technology of Segment Anything Model (SAM). Specific improvements include: - **Enhanced Attention Dropout Layer (Enhanced ADL)**: By explicit visual prompting, it solves the problems of partial activation and mis - activation in CAM, thereby generating more accurate initial pseudo - labels. - **Pixel - level Entropy - based Prompt Module (PEPM)**: It uses high - entropy points as prompts to improve the performance of SAM under weakly - supervised conditions. - **Iterative retraining strategy**: By initially fine - tuning the mask decoder of SAM and selecting reliable pseudo - labels for iterative retraining, the model performance is gradually optimized. ### Experimental results The experimental results show that the EP - SAM method outperforms the existing weakly - supervised and fully - supervised methods on three breast cancer pathological image datasets. Especially on the Camelyon17 and Camelyon16 datasets, it even surpasses fully - supervised models such as MedSAM and U - Net. ### Summary This paper proposes a new weakly - supervised pathological image segmentation method by combining the advantages of CAM and SAM, effectively solves the problem of insufficient labeled data, and shows superior performance on multiple datasets.

EP-SAM: Weakly Supervised Histopathology Segmentation via Enhanced Prompt with Segment Anything

Weakly supervised semantic segmentation of histological tissue via attention accumulation and pixel-level contrast learning

SAM-Path: A Segment Anything Model for Semantic Segmentation in Digital Pathology

Path-SAM2: Transfer SAM2 for digital pathology semantic segmentation

Guided Prompting in SAM for Weakly Supervised Cell Segmentation in Histopathological Images

WSI-SAM: Multi-resolution Segment Anything Model (SAM) for histopathology whole-slide images

Weakly Supervised Histopathology Image Segmentation with Sparse Point Annotations

Morphology-Enhanced CAM-Guided SAM for weakly supervised Breast Lesion Segmentation

Evaluation and improvement of Segment Anything Model for interactive histopathology image segmentation

Weakly-Supervised Semantic Segmentation for Histopathology Images Based on Dataset Synthesis and Feature Consistency Constraint.

Multi-layer pseudo-supervision for histopathology tissue semantic segmentation using patch-level classification labels

CellSAM: Advancing Pathologic Image Cell Segmentation via Asymmetric Large‐Scale Vision Model Feature Distillation Aggregation Network

HAMIL: High-Resolution Activation Maps and Interleaved Learning for Weakly Supervised Segmentation of Histopathological Images

Pathological Primitive Segmentation Based on Visual Foundation Model with Zero-Shot Mask Generation

A robust image segmentation and synthesis pipeline for histopathology

CAMEL: A Weakly Supervised Learning Framework for Histopathology Image Segmentation

WeakSAM: Segment Anything Meets Weakly-supervised Instance-level Recognition

A Comprehensive Analysis of Weakly-Supervised Semantic Segmentation in Different Image Domains

Effect of diet and tylosin on chronic diarrhea in beagles.

A randomized multicenter trial comparing leukocyte function-associated antigen-1 monoclonal antibody with rabbit antithymocyte globulin as induction treatment in first kidney transplantations.

Online Easy Example Mining for Weakly-supervised Gland Segmentation from Histology Images