International Journal of Information Technology and Applications, Vol. 1, No. 1, pp. 1-8, March 2024.
Abstract: Gastroscopic image enhancement is an important research direction in the field of medical image processing, which aims at improving the image quality, and enhancing the accuracy of clinical diagnosis. However, existing algorithms are generally designed with large-scale deep learning network structures, which require high computing resources and storage space, and thus it is difficult to achieve their practical application and scalability. In this paper, we proposed a lightweight network for gastroscopic image enhancement (LNet-GIE). The LNet-GIE first loads the pre-trained weights for initialization, and then passes them into the constructed lightweight network, only designed with common convolutional layers. Quantitative and qualitative experiments on public gastroscopic image enhancement dataset show that the size of LNet-GIE is significantly smaller, while the performances are better than the state-of-the-art methods. The code of the proposed LGNet-GIE is public and available at https://github.com/20618xx/LNet-GIE.