Research

Paper

TESTING March 18, 2026

ResNet-50 with Class Reweighting and Anatomy-Guided Temporal Decoding for Gastrointestinal Video Analysis

Authors

Romil Imtiaz, Dimitris K. Iakovidis

Abstract

We developed a multi-label gastrointestinal video analysis pipeline based on a ResNet-50 frame classifier followed by anatomy-guided temporal event decoding. The system predicts 17 labels, including 5 anatomy classes and 12 pathology classes, from frames resized to 336x336. A major challenge was severe class imbalance, particularly for rare pathology labels. To address this, we used clipped class-wise positive weighting in the training loss, which improved rare-class learning while maintaining stable optimization. At the temporal stage, we found that direct frame-to-event conversion produced fragmented mismatches with the official ground truth. The final submission therefore combined GT-style framewise event composition, anatomy vote smoothing, and anatomy-based pathology gating with a conservative hysteresis decoder. This design improved the final temporal mAP from 0.3801 to 0.4303 on the challenge test set.

Metadata

arXiv ID: 2603.17784
Provider: ARXIV
Primary Category: cs.CV
Published: 2026-03-18
Fetched: 2026-03-19 06:01

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.17784v1</id>\n    <title>ResNet-50 with Class Reweighting and Anatomy-Guided Temporal Decoding for Gastrointestinal Video Analysis</title>\n    <updated>2026-03-18T14:46:40Z</updated>\n    <link href='https://arxiv.org/abs/2603.17784v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.17784v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>We developed a multi-label gastrointestinal video analysis pipeline based on a ResNet-50 frame classifier followed by anatomy-guided temporal event decoding. The system predicts 17 labels, including 5 anatomy classes and 12 pathology classes, from frames resized to 336x336. A major challenge was severe class imbalance, particularly for rare pathology labels. To address this, we used clipped class-wise positive weighting in the training loss, which improved rare-class learning while maintaining stable optimization. At the temporal stage, we found that direct frame-to-event conversion produced fragmented mismatches with the official ground truth. The final submission therefore combined GT-style framewise event composition, anatomy vote smoothing, and anatomy-based pathology gating with a conservative hysteresis decoder. This design improved the final temporal mAP from 0.3801 to 0.4303 on the challenge test set.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.LG'/>\n    <published>2026-03-18T14:46:40Z</published>\n    <arxiv:comment>ICPR 2026 RARE-VISION Competition</arxiv:comment>\n    <arxiv:primary_category term='cs.CV'/>\n    <author>\n      <name>Romil Imtiaz</name>\n    </author>\n    <author>\n      <name>Dimitris K. Iakovidis</name>\n    </author>\n  </entry>"
}