Automatic Polyp Segmentation Using Convolutional Neural Networks
This work addresses the need for improved polyp detection in colonoscopies to reduce missed diagnoses, but it is incremental as it compares existing architectures without introducing a novel method.
The paper tackled the problem of automatic polyp segmentation to aid in early colorectal cancer diagnosis by comparing deep learning architectures as feature extractors in a U-Net framework, with the DenseNet169-based model achieving an accuracy of 99.15%, Dice coefficient of 90.87%, and Jaccard index of 83.82% on the CVC-Clinic dataset.
Colorectal cancer is the third most common cancer-related death after lung cancer and breast cancer worldwide. The risk of developing colorectal cancer could be reduced by early diagnosis of polyps during a colonoscopy. Computer-aided diagnosis systems have the potential to be applied for polyp screening and reduce the number of missing polyps. In this paper, we compare the performance of different deep learning architectures as feature extractors, i.e. ResNet, DenseNet, InceptionV3, InceptionResNetV2 and SE-ResNeXt in the encoder part of a U-Net architecture. We validated the performance of presented ensemble models on the CVC-Clinic (GIANA 2018) dataset. The DenseNet169 feature extractor combined with U-Net architecture outperformed the other counterparts and achieved an accuracy of 99.15\%, Dice similarity coefficient of 90.87%, and Jaccard index of 83.82%.