کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

DepthFusionSAM: Training-Free Colonoscopy Polyp Segmentation via Retrieval-Guided Prompting

نویسندگان :

Ali Tamizifar¹ Shakiba Berenjkoub² Nader Karimi³ Shadrokh Samavi⁴

1- Department of Electrical and Computer Engineering, Isfahan University of Technology 2- Department of Electrical and Computer Engineering, Isfahan University of Technology 3- Department of Electrical and Computer Engineering, Isfahan University of Technology 4- Department of Computer Science, Seattle University

کلمات کلیدی :

Polyp segmentation،Colonoscopy،SAM3،Training-free learning،SegGPT،DepthFusionSAM

چکیده :

Accurate colorectal polyp segmentation is vital for early cancer detection, yet existing deep learning methods often require extensive training or fine-tuning to handle the high variability and complex backgrounds of colonoscopy images. This paper introduces DepthFusionSAM, a novel three-stage, training-free framework for automated polyp segmentation. The framework first utilizes DINOv2 embeddings and FAISS models to retrieve visually similar template images that provide query-specific priors. In the second stage, SegGPT generates multiple candidate masks, which are then re-ranked using semantic consistency and scale-based priors. Finally, a median box prompt is fed into SAM3, augmented by DepthFusion—a depth-guided feature gating mechanism derived from a pre-trained depth estimation model—to enhance robustness against specular highlights and low-contrast boundaries. Evaluated on benchmark datasets including Kvasir-SEG and PolypGen, DepthFusionSAM achieves competitive performance without any dataset-specific optimization, offering a scalable and efficient solution for real-time clinical workflows.

لیست مقالات

لیست مقالات بایگانی شده

A New Unsupervised Feature Learning Method for Object Recognition using Prior-Knowledge Data

Ashkan Farrokhi - Hadi Seyedarabi

Ultra-Low Power Current-Mode ASK Demodulator for Contactless Smart Cards

Somayeh Yousefi - Mohsen Jalali

Design and fabrication tip tapered fiber optic dopamine sensor based on LSPR

Roksana Esmaeilpour - Mohammad Ismail zibaii - Masoumeh Barkand - Marzieh Pajouhandeh - Soroush Rostami - Mehdi Banihashemi - Mohammad-Mahdi Babakhani-fard

Sensitive RSNs to Schizophrenia; A graph parameter approach

Shirin Karimian - Farzaneh Keyvanfard - Abbas Nasiraei Moghaddam

Giant Optical Nonreciprocity with Magnetized Epsilon-Near-Zero Materials

Zahra Chamani - Abolghasem Zeidaabadi Nezhad - Mahyar Dehdast - Zaker Hossein Firouzeh

Entanglement-Assisted Classical-Quantum Multiple Access Wiretap Channel: One-shot Achievable Rate Region

Hadi Aghaee - Bahareh Akhbari

A Two-Step Stochastic Market-Oriented Approach for Optimal Operation of Commercial VPPs under Uncertainty

Jalal Moradi - Hossein Shahinzadeh - Ahmad Hafezimagham - Gevork B. Gharehpetian - S.M. Muyeen - Mohamed Benbouzid

یک مدل آگاهی موقعیتی برای تشخیص بحران در شبکه‌های حسگر بی‌سیم

مهدی سالخورده حقیقی - امیرمحمد رضوانی نیا

ساخت حسگر مقاومتی بخار اتانول مبتنی بر هتروساختار باریم تیتانات / اکسید روی آلاییده با نانوذرات نقره

محسن طاهری پور - نوید یثربی - شیرین نصراصفهانی - محمد حسین شیخی

Cost- and Emission-Optimized Energy Management in Microgrids Integrating Renewables, Storage, and Electric Vehicle

Mahdi Aghamohammadi Gharabagh - Tohid Ghanizadeh Bolandi - Behrouz Tousi

بیشتر