Modified Snapshot Ensemble Algorithm for Skin Lesion Classification

会议论文
发表日期：

2024年

查看全文

Samson Anosh Babu Parisapogu,
Mastan Mohammed Meera Durga,
Vallela Kaushik Shashank Reddy,
Boyapati Kalyan Chakravarthi,
P. Vasanth Sena

摘要

Skin cancer prediction has become an essential task in dermoscopic image analysis. For automatic diagnosis of the skin lesions, the Scientific community. The most crucial part in the cure of skin cancer is the exact identification and classification of the skin lesion types. This paper proposes a modified snapshot ensemble algorithm for skin lesion classification. This method utilizes the advantages of transfer learning approach with ResNet50. The method is shown efficient results for the popular metrics, such as accuracy, F1-score, recall and precision.