LPNet: A lightweight CNN with discrete wavelet pooling strategies for colon polyps classification

• DOI: 10.1002/ima.22825
• Publication Date: 2022-11-14T11:06:50Z
• Authors: Pallabi Sharma; Dipankar Das; Anmol Gautam; Bunil Kumar Balab...
• Keywords: Pooling

Abstract The traditional process of disease diagnosis from medical images follows a manual process, which is tedious and arduous. A computer‐aided (CADs) system can work as an assistive tool to improve the process. In this pursuit, article introduces unique architecture LPNet for classifying colon polyps colonoscopy video frames. Colon are abnormal growth cells in wall. Over time, untreated may cause colorectal cancer. Different convolutional neural networks (CNNs) based systems have been developed recent years. However, CNN uses pooling reduce number parameters expand receptive field. On other hand, results data loss deleterious subsequent processes. Pooling strategies on discrete wavelet operations proposed our solution problem, with promise achieving better trade‐off between field size computing efficiency. overall performance model superior others, according experimental dataset. bio‐orthogonal achieved highest accuracy 93.55%. It outperforms state‐of‐the‐art (SOTA) models classification task, it lightweight terms learnable compared them, making easily deployable edge devices.