报告详情
RaP-ProtoViT: Efficient Dual-Head Transformers for Robust Gastric Endoscopy Classification and Generalizable Clinical Deployment
编号：164
访问权限：仅限参会人
更新：2025-12-23 13:29:05 浏览：3次
拓展类型2
报告开始：暂无开始时间（Asia/Amman）
报告时间：暂无持续时间
所在会场：[暂无会议] [暂无会议段]
暂无文件
摘要
We introduce RaP-ProtoViT, an end-to-end dual-head transformer for 8-class GI endoscopy (Kvasir-v2). A margin head (ArcFace/AM-Softmax) enforces angular separation, while a prototype head aggregates top-k token–prototype similarities (with M trainable prototypes/class); a lightweight input-adaptive MLP fuses the heads. A leakage-aware pipeline (pHash dedup + GroupKFold) prevents near-duplicate bleed-over. Training uses AdamW(+SAM) with cosine warm-up, DropPath, label smoothing, SWA, and post-hoc temperature scaling; two-stage HPO (MOTPE+ASHA → qEHVI) under Latency@224 ≤ 200 ms and memory constraints selects operating points. On Kvasir-v2 the model attains 99.1% accuracy, Macro-F1 = 0.991, Macro-AUPRC = 0.997, AUROC = 0.998, and ECE ≈ 0.9%, with per-class F1 tightly clustered in 0.988–0.994 and fold stability (±0.2 pp accuracy, ±0.002 Macro-F1). Ablations show margin-only/prototype-only variants reduce Macro-F1 to 0.967/0.975 and raise ECE to 2.8%/2.2%; removing adaptive fusion drops Macro-F1 to 0.984. The proposed HPO converges 2–3× faster and yields better final MF1/AUPRC/ECE than Bayesian TPE or Random+ASHA. The prototype head provides localized, intrinsically interpretable evidence, complementing the margin head’s discrimination, within a single-model deployment footprint. By advancing robust, interpretable, and computationally efficient AI for gastric endoscopy, our approach can improve early detection of gastrointestinal disease and enable reliable clinical deployment across diverse healthcare settings.
关键词
Endoscopy classification, Vision transformer, Prototype learning, hyperparameter optimization.
报告人
Mohamadreza Khosravi
Researcher Shiraz University of Medical Sciences稿件作者
全部评论
重要日期
-
会议日期
12月29日
2025
至12月31日
2025
-
12月30日 2025
报告提交截止日期
-
12月30日 2025
注册截止日期
-
12月31日 2025
初稿截稿日期
主办单位
国际科学联合会
承办单位
扎尔卡大学
联系方式
- Miss AsianComNet
- as******@usssociety.org
历届会议
-
2024年10月24日 泰国 Bangkok
2024亚洲通讯与网络会议
发表评论