Article citationsMore>>

Wisdom ML (2024) Understanding ResNet-50 in Depth: Architecture, Skip Connections, and Advantages over Other Networks.
https://wisdomml.in/understanding-resnet-50-in-depth-architecture-skip-connections-and-advantages-over-other-networks

has been cited by the following article:

• TITLE: Deep Learning Approaches for the Identification and Classification of Skin Cancer

  AUTHORS: Kanchon Kumar Bishnu, Mohammad Abu Saleh, Saddam Hossain, Jannatul Ferdous Mou, Mia Md. Tofayel Gonee Manik, Araf Islam

  KEYWORDS: ResNet50, Convolutional Neural Network, Deep Learning, Augmentation, Preprocessing

  JOURNAL NAME: Journal of Computer and Communications, Vol.12 No.12, December 19, 2024

  ABSTRACT: One of the most dangerous forms of cancer, skin cancer has been on the rise over the past ten years. Nonetheless, melanoma detection is a method that uses deep learning algorithms to analyze images and accurately diagnose melanoma. An improved result for cancer treatment may result from early diagnosis. Then, in a matter of seconds, it will be simple to identify skin cancer using deep learning. In this research, a deep learning-based automatic skin cancer detection method is proposed. Data was considered from the ISIC database dataset which has 2357 images. To obtain average color information and normalize all color channel information, we used a few preprocessing approaches. Next, data was collected for categorization and reshaping of the images. To avoid overfitting, we additionally employed data augmentation. In the end, the Convolutional Neural Network was used to achieve our goal, which improved the accuracy of prediction. Using the Resnet50 algorithm, the accuracy rate rose to 98%, which will be helpful to get a good outcome with better accuracy.