CNL-UNet: A novel lightweight deep learning architecture for multimodal biomedical image segmentation with false output suppression

Md. Badiuzzaman Shuvo, Rifat Ahommed, Sakib Reza, M. M. A. Hashem

Abstract: Highlights•A novel lightweight architecture is proposed based on a U-shaped architecture.•A novel CNL module is introduced for suppressing false positive and negative outputs.•Res-path is used for better semantic compatibility between the encoder and decoder.•The concept of transfer learning is adapted to speed up the convergence procedure.•Experimental results show that the proposed CNL-UNet outperforms existing networks.