AI Paper Atlas

Research map / Detection and segmentation / Weakly supervised detection and segmentation

CAMEL: A Weakly Supervised Learning Framework for Histopathology Image Segmentation

Gang Xu, Zhigang Song, Zhuo Sun, Calvin Ku, Zhe Yang, Cancheng Liu, Shuhao Wang, Jianpeng Ma, Wei Xu

ICCV 2019 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised semantic segmentation · 197 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
combine Max-Max and Max-Min selection (cMIL) instead of Max-Max alone (ablation)instance classifier (320) · CAMELYON16 testaccuracy: 81.9 -> 92.3Improved
combine Max-Max and Max-Min selection (cMIL) instead of Max-Min alone (ablation)instance classifier (320) · CAMELYON16 testaccuracy: 82.3 -> 92.3Improved
combine Max-Max and Max-Min selection (cMIL) instead of Max-Max alone (ablation)instance classifier (160) · CAMELYON16 testaccuracy: 79.3 -> 88.4Improved
add image-level constraints to retraining (ablation)instance classifier (320) · CAMELYON16 testaccuracy: 92.3 -> 92.9Improved
add image-level constraints to retraining (ablation)instance classifier (160) · CAMELYON16 testaccuracy: 88.4 -> 89.9Improved
add cascade data enhancement (ablation)instance classifier (160) · CAMELYON16 testaccuracy: 88.4 -> 90.4Improved
add cascade data enhancement to the constrained model (ablation)instance classifier (160) · CAMELYON16 testaccuracy: 89.9 -> 91.7Improved
use enriched instance labels instead of image-level labels assigned to pixelsDeepLabv2 · CAMELYON16 testIoU: 72.9 -> 85.4Improved
use enriched instance labels instead of image-level labels assigned to pixelsU-Net · CAMELYON16 testIoU: 70.6 -> 84.8Improved
use finer instances (160 instead of 320) (ablation)CAMEL DeepLabv2 · CAMELYON16 testIoU: 84.3 -> 85.4Improved
use CAMEL instead of CDWS-MILsegmentation · CAMELYON16 testIoU: 77.6 -> 85.4Improved
use finer instances (160 instead of 320) (ablation)CAMEL DeepLabv2 · Colorectal adenomaF1-score: 86.1 -> 87.3Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy