Impact Statement:Esophageal cancer is one of the most common causes of cancer deaths worldwide. Early diagnosis can significantly reduce the mortality rate of esophageal cancer. Artificia...Show More
Abstract:
Automatic esophageal lesion identification (ESEI) is of great importance to clinically aid the endoscopists with the early detection of esophageal cancer. However, accura...Show MoreMetadata
Impact Statement:
Esophageal cancer is one of the most common causes of cancer deaths worldwide. Early diagnosis can significantly reduce the mortality rate of esophageal cancer. Artificial intelligence (AI) methods, particularly neural networks, have been extensively utilized to develop automatic esophageal lesion systems, which aims at assisting the endoscopists during gastroscopic screening. Existing studies manually design a valid neural network architecture for esophageal lesion identification which requires high expertise and a large workload. The evolutionary neural architecture search algorithm introduced in this article designs a multitask network search space and further incorporates a one-shot supernet strategy in the evolutionary algorithm for searching the optimal network architecture. The approach has the potential to alleviate the burden of neural network architecture design while improving the performance for esophageal lesion identification under clinical scenarios.
Abstract:
Automatic esophageal lesion identification (ESEI) is of great importance to clinically aid the endoscopists with the early detection of esophageal cancer. However, accurate identification of esophageal lesion is challenging due to the varying shape, size, illumination condition, and complex background with artifacts in endoscopic images. Although deep neural network based approaches have considerably boosted the performance by automatically learning features from esophageal images, the configuration of the network architecture is highly dependent on domain expertise and is a daunting task to be manually tuned. In this article, we propose an evolutionary algorithm based approach to search for the optimal multitask network architecture for ESEI. Different from existing studies, we first design a multitask network search space, which considers the lesion identification as two steps including esophageal image classification and esophageal lesion segmentation. In particular, the input image...
Published in: IEEE Transactions on Artificial Intelligence ( Volume: 3, Issue: 3, June 2022)