Narrow Results

Dermatologic oncology’s precision medicine revolutionizes skin cancer detection by integrating advanced technologies and personalized patient data. Dermatologic oncology concentrates on detecting skin cancer, utilizing modern techniques and technologies to detect and classify several kinds of cutaneous malignancies. Leveraging medical knowledge or advanced imaging approaches like dermoscopy and reflectance confocal microscopy; dermatologists effectively investigate skin cancer for subtle signs of malignancy. Furthermore, computer-aided diagnostic (CAD) systems, controlled by machine learning (ML) methods, are gradually deployed to boost diagnostic accuracy by investigating massive datasets of dermatoscopic images. It is a multi-disciplinary method that allows early recognition of skin lesions and enables precise prognostication and particular treatment approach, finally enhancing patient outcomes in dermatologic oncology. This paper presents the Fractals Snake Optimization with Deep Learning for Accurate Classification of Skin Cancer in Dermoscopy Images (SODL-ACSCDI) approach. The purpose of the SODL-ACSCDI approach is to identify and categorize the existence of skin cancer on Dermoscopic images. The SODL-ACSCDI technique applies a contrast enhancement process as the initial step. Next, the SODL-ACSCDI technique involves the SE-ResNet+FPN model for deriving intrinsic and complex feature patterns from dermoscopic images. Additionally, the SO technique can help boost the hyperparameter selection of the SE-ResNet+FPN approach. Furthermore, skin cancer classification uses the convolutional autoencoder (CAE) approach. The experimentation results of the SODL-ACSCDI technique could be examined using a dermoscopic image dataset. A wide-ranging result of the SODL-ACSCDI technique indicated a superior performance of 99.61% compared to recent models concerning various metrics.

Deep learning-based skin lesion segmentation methods have achieved promising results in the community. However, they are usually based on fully supervised learning and require many high-quality ground truths. Labeling the ground truths takes a lot of labor, material, and financial resources. We propose a novel semi-supervised skin lesion segmentation method to solve this problem. First, a hierarchical image segmentation algorithm is used to generate optimal segmentation maps. Then, fully supervised training is performed on a small part of the images with ground truths. The resulting pseudo masks are generated to train the rest of the images. The optimal segmentation maps are utilized in this process to refine the pseudo masks. Experiments show that the proposed method can improve the performance of semi-supervised learning for skin lesion segmentation by reducing the gap with fully supervised learning methods. Moreover, it can reduce the workload of labeling the ground truths. Extensive experiments are conducted on the open dataset to validate the efficiency of the proposed method. The results show that our method is competitive in improving the quality of semi-supervised segmentation.