Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Nguyen, Y Hop, Huu, Doan Anh Phan, Tran, Trung Thai, Mai, Nhat Nam, Giap, Van Toi, Dao, Thao Thi Phuong, Le, Trung-Nghia
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2509.00752
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915472499277824
author Nguyen, Y Hop
Huu, Doan Anh Phan
Tran, Trung Thai
Mai, Nhat Nam
Giap, Van Toi
Dao, Thao Thi Phuong
Le, Trung-Nghia
author_facet Nguyen, Y Hop
Huu, Doan Anh Phan
Tran, Trung Thai
Mai, Nhat Nam
Giap, Van Toi
Dao, Thao Thi Phuong
Le, Trung-Nghia
contents We present a unified vision-language framework tailored for ENT endoscopy image analysis that simultaneously tackles three clinically-relevant tasks: image classification, image-to-image retrieval, and text-to-image retrieval. Unlike conventional CNN-based pipelines that struggle to capture cross-modal semantics, our approach leverages the CLIP ViT-B/16 backbone and enhances it through Low-Rank Adaptation, multi-level CLS token aggregation, and spherical feature interpolation. These components collectively enable efficient fine-tuning on limited medical data while improving representation diversity and semantic alignment across modalities. To bridge the gap between visual inputs and textual diagnostic context, we introduce class-specific natural language prompts that guide the image encoder through a joint training objective combining supervised classification with contrastive learning. We validated our framework through participation in the ACM MM'25 ENTRep Grand Challenge, achieving 95% accuracy and F1-score in classification, Recall@1 of 0.93 and 0.92 for image-to-image and text-to-image retrieval respectively, and MRR scores of 0.97 and 0.96. Ablation studies demonstrated the incremental benefits of each architectural component, validating the effectiveness of our design for robust multimodal medical understanding in low-resource clinical settings.
format Preprint
id arxiv_https___arxiv_org_abs_2509_00752
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Multi-Level CLS Token Fusion for Contrastive Learning in Endoscopy Image Classification
Nguyen, Y Hop
Huu, Doan Anh Phan
Tran, Trung Thai
Mai, Nhat Nam
Giap, Van Toi
Dao, Thao Thi Phuong
Le, Trung-Nghia
Computer Vision and Pattern Recognition
We present a unified vision-language framework tailored for ENT endoscopy image analysis that simultaneously tackles three clinically-relevant tasks: image classification, image-to-image retrieval, and text-to-image retrieval. Unlike conventional CNN-based pipelines that struggle to capture cross-modal semantics, our approach leverages the CLIP ViT-B/16 backbone and enhances it through Low-Rank Adaptation, multi-level CLS token aggregation, and spherical feature interpolation. These components collectively enable efficient fine-tuning on limited medical data while improving representation diversity and semantic alignment across modalities. To bridge the gap between visual inputs and textual diagnostic context, we introduce class-specific natural language prompts that guide the image encoder through a joint training objective combining supervised classification with contrastive learning. We validated our framework through participation in the ACM MM'25 ENTRep Grand Challenge, achieving 95% accuracy and F1-score in classification, Recall@1 of 0.93 and 0.92 for image-to-image and text-to-image retrieval respectively, and MRR scores of 0.97 and 0.96. Ablation studies demonstrated the incremental benefits of each architectural component, validating the effectiveness of our design for robust multimodal medical understanding in low-resource clinical settings.
title Multi-Level CLS Token Fusion for Contrastive Learning in Endoscopy Image Classification
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.00752