Rethinking the transfer learning for FCN based polyp segmentation in colonoscopy
This work addresses the problem of improving polyp segmentation accuracy for medical imaging applications, representing an incremental advance with specific gains in a domain-specific context.
The paper tackled the challenge of automated polyp segmentation in colonoscopy images, which suffers from small and imbalanced datasets leading to overfitting, by proposing a pipeline that couples segmentation and classification tasks, achieving improvements of 4.34% and 5.70% in Polyp-IoU on two datasets compared to state-of-the-art methods.
Besides the complex nature of colonoscopy frames with intrinsic frame formation artefacts such as light reflections and the diversity of polyp types/shapes, the publicly available polyp segmentation training datasets are limited, small and imbalanced. In this case, the automated polyp segmentation using a deep neural network remains an open challenge due to the overfitting of training on small datasets. We proposed a simple yet effective polyp segmentation pipeline that couples the segmentation (FCN) and classification (CNN) tasks. We find the effectiveness of interactive weight transfer between dense and coarse vision tasks that mitigates the overfitting in learning. And It motivates us to design a new training scheme within our segmentation pipeline. Our method is evaluated on CVC-EndoSceneStill and Kvasir-SEG datasets. It achieves 4.34% and 5.70% Polyp-IoU improvements compared to the state-of-the-art methods on the EndoSceneStill and Kvasir-SEG datasets, respectively.