Synthetic Data-Driven Multi-Architecture Framework for Automated Polyp Segmentation Through Integrated Detection and Mask Generation

TL;DR

This paper presents a multi-architecture framework combining synthetic data generation, detection, and segmentation to improve automated polyp segmentation in colonoscopy images, addressing dataset limitations and annotation challenges.

Contribution

It introduces a novel integrated system using synthetic data, Faster R-CNN, and SAM for enhanced polyp detection and segmentation, with comprehensive evaluation of multiple segmentation models.

Findings

Faster R-CNN achieved 93.08% recall and 88.97% precision.

FPN outperformed other models with highest PSNR and SSIM.

UNet had the best recall among segmentation models.

Abstract

Colonoscopy is a vital tool for the early diagnosis of colorectal cancer, which is one of the main causes of cancer-related mortality globally; hence, it is deemed an essential technique for the prevention and early detection of colorectal cancer. The research introduces a unique multidirectional architectural framework to automate polyp detection within colonoscopy images while helping resolve limited healthcare dataset sizes and annotation complexities. The research implements a comprehensive system that delivers synthetic data generation through Stable Diffusion enhancements together with detection and segmentation algorithms. This detection approach combines Faster R-CNN for initial object localization while the Segment Anything Model (SAM) refines the segmentation masks. The faster R-CNN detection algorithm achieved a recall of 93.08% combined with a precision of 88.97% and an F1 score of 90.98%.SAM is then used to generate the image mask. The research evaluated five state-of-the-art segmentation models that included U-Net, PSPNet, FPN, LinkNet, and MANet using ResNet34 as a base model. The results demonstrate the superior performance of FPN with the highest scores of PSNR (7.205893) and SSIM (0.492381), while UNet excels in recall (84.85%) and LinkNet shows balanced performance in IoU (64.20%) and Dice score (77.53%).