脑瘤
分割
人工智能
假阳性悖论
计算机科学
医学影像学
规范化(社会学)
模式识别(心理学)
空间归一化
特征(语言学)
医学
图像分割
计算机视觉
目标检测
特征提取
脑病
过程(计算)
磁共振成像
图像处理
神经影像学
文本挖掘
阶段(地层学)
对比度增强
放射科
医学诊断
人脑
对比度(视觉)
作者
Premnath K.,S. Pandiaraj
出处
期刊:
[Institute of Electrical and Electronics Engineers]
日期:2025-09-24
卷期号:: 1-8
标识
DOI:10.1109/wccest66994.2025.11389963
摘要
In order to make a proper diagnosis of brain tumor at the early stages of the disease, it is necessary to have precise information concerning the location and size of tumors as well as the information about the type of tumor and the disease occurrence. Human interpretation of MRI is still not feasible given that it takes too much processing time and is characterized by random human error that renders the process unfavorable in medical diagnostics. The study introduces a deep learning-based system of brain tumor detection to obtain greater detection rates and performance. The article gives a comparison of the YOLO based segmentation versus CNN classification in the detection of brain tumors and their location within MRI images. Researchers evaluated their system on many cases of brain tumor images with various tumor size and shape and contrast of intensity. The model performance is also improved by the normalization skills and data augmentation processes. YOLO v8 real time object detection has a higher accuracy in tumor feature extraction since it reduces the false positives needed to provide medical diagnostic accuracy. Deep learning-based detection models have been found to be beneficial to medical image and clinical decision and have demonstrated strong gains when applied in tumor detection experiments hence reflecting a bright future application in the field. By means of these technological developments, medical practitioners are able to attain greater accuracy in the diagnosis area besides enhanced curative techniques to attain improved health care outcomes since they can be used to enhance therapeutic strategies.
科研通智能强力驱动
Strongly Powered by AbleSci AI