MFSNet: A multi focus segmentation network for skin lesion segmentation

Segmentation is essential for medical image analysis to identify and localize diseases, monitor morphological changes, extract discriminative features further diagnosis. Skin cancer one of the most common types globally, its early diagnosis pivotal complete elimination malignant tumors from body. This research develops an Artificial Intelligence (AI) framework supervised skin lesion segmentation employing deep learning approach. The proposed framework, called MFSNet (Multi-Focus Network), uses differently scaled feature maps computing final mask using raw input RGB images lesions. In doing so, initially, are preprocessed remove unwanted artifacts noises. employs Res2Net backbone, a recently convolutional neural network (CNN), obtaining used in Parallel Partial Decoder (PPD) module get global map mask. different stages network, convolution multi-scale two boundary attention (BA) modules reverse (RA) generate output. MFSNet, when evaluated on three publicly available datasets: $PH^2$, ISIC 2017, HAM10000, outperforms state-of-the-art methods, justifying reliability framework. relevant codes approach accessible at https://github.com/Rohit-Kundu/MFSNet