An Efficient Polyp Segmentation Network

TL;DR

This paper introduces a lightweight deep learning model for polyp segmentation that outperforms existing models in accuracy while using significantly fewer parameters, aiding early colon cancer detection.

Contribution

A novel polyp segmentation network combining EfficientNetB0 encoder, partial decoder, asymmetric convolution, and squeeze-and-excitation blocks, with fewer parameters and improved accuracy.

Findings

Achieved higher Dice scores than state-of-the-art models.

Reduced parameter count to approximately 2.6 million.

Outperformed existing models on multiple datasets.

Abstract

Cancer is a disease that occurs as a result of the uncontrolled division and proliferation of cells. Colon cancer is one of the most common types of cancer in the world. Polyps that can be seen in the large intestine can cause cancer if not removed with early intervention. Deep learning and image segmentation techniques are used to minimize the number of polyps that goes unnoticed by the experts during these interventions. Although these techniques perform well in terms of accuracy, they require too many parameters. We propose a new model to address this problem. Our proposed model requires fewer parameters as well as outperforms the state-of-the-art models. We use EfficientNetB0 for the encoder part, as it performs well in various tasks while requiring fewer parameters. We use partial decoder, which is used to reduce the number of parameters while achieving high accuracy in segmentation. Since polyps have variable appearances and sizes, we use an asymmetric convolution block instead of a classic convolution block. Then, we weight each feature map using a squeeze and excitation block to improve our segmentation results. We used different splits of Kvasir and CVC-ClinicDB datasets for training, validation, and testing, while we use CVC- ColonDB, ETIS, and Endoscene datasets for testing. Our model outperforms state-of-art models with a Dice metric of %71.8 on the ColonDB test dataset, %89.3 on the EndoScene test dataset, and %74.8 on the ETIS test dataset while requiring fewer parameters. Our model requires 2.626.337 parameters in total while the closest model in the state-of-the-art is U-Net++ with 9.042.177 parameters.