Scholarlink

A brain tumor is an aberrant development of brain cells that can be benign or malignant. Radiation exposure, environmental conditions, and genetic abnormalities are among the causes of brain tumors. Effective therapy and improved patient outcomes depend on early identification and precise categorization. In medical image analysis, deep learning has demonstrated great potential for the classification of brain cancers. Convolutional neural networks (CNNs), one type of deep learning model, automatically extract relevant features from raw data, in contrast to traditional approaches that rely on manually constructed features. Through the use of self-attention processes, Vision Transformers (ViTs) have recently demonstrated potential in medical imaging tasks by capturing long-range dependencies in pictures. To categorize tumors, ViTs split an image into patches, encode them, and then apply transformer layers and analytics based on different batch sizes to get better performance of ViT. In the proposed study different ViTs are used on different batch sizes and it will give accuracy around near to 98% which is better than traditional classification models.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Batch Size Optimization in Vision Transformer-Based Brain Tumor Classification

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Kshitij Oza,
  • Ronak Patel,
  • Dhruvil Joshi,
  • Miral Patel

摘要

A brain tumor is an aberrant development of brain cells that can be benign or malignant. Radiation exposure, environmental conditions, and genetic abnormalities are among the causes of brain tumors. Effective therapy and improved patient outcomes depend on early identification and precise categorization. In medical image analysis, deep learning has demonstrated great potential for the classification of brain cancers. Convolutional neural networks (CNNs), one type of deep learning model, automatically extract relevant features from raw data, in contrast to traditional approaches that rely on manually constructed features. Through the use of self-attention processes, Vision Transformers (ViTs) have recently demonstrated potential in medical imaging tasks by capturing long-range dependencies in pictures. To categorize tumors, ViTs split an image into patches, encode them, and then apply transformer layers and analytics based on different batch sizes to get better performance of ViT. In the proposed study different ViTs are used on different batch sizes and it will give accuracy around near to 98% which is better than traditional classification models.