Scholarlink

Benign brain tumors result from abnormal cell growth within the brain. The death rates can’t be established because the disease is rare and has many classifications in its ambit. MRI scans are greatly valid in finding tumors (Shah et al in IEEE Access, vol. 10, pp. 65426–65438 (2022)). But the procedure relating to finding tumors in images is manual. Hence, this approach is time-consuming and may lead to variable results depending on the radiologist’s experience. These are the limitations that become important to overcome. The uncontrollable advancements in the field of artificial intelligence are developing especially computer-aided methods (Shah et al. in IEEE Access, vol. 10, pp. 65426–65438 (2022)). This research proposes a deep complex neural network model, namely, advanced semantic segmentation derived from an efficient B0 network for correct identification and detection of brain tumors from MRI images. Image enhancement techniques were employed to improve image quality and training data variability. With the use of enhancement techniques, the size increases. The other DL models included in the comparative analysis are VGG16, InceptionV3, Xception, ResNet50, and InceptionResNetV2.

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

Brain Tumor Detection Using Deep Learning with EfficientNet-B0

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Chandana Yamani,
  • Sireesha Moturi,
  • Devarasetty Rama Durga Bhavani,
  • Bollisetty Triveni,
  • Pondugula Venkata Naga Hemantha Lakshmi,
  • Dodda Venkata Reddy,
  • S. Siva Nageswara Rao

摘要

Benign brain tumors result from abnormal cell growth within the brain. The death rates can’t be established because the disease is rare and has many classifications in its ambit. MRI scans are greatly valid in finding tumors (Shah et al in IEEE Access, vol. 10, pp. 65426–65438 (2022)). But the procedure relating to finding tumors in images is manual. Hence, this approach is time-consuming and may lead to variable results depending on the radiologist’s experience. These are the limitations that become important to overcome. The uncontrollable advancements in the field of artificial intelligence are developing especially computer-aided methods (Shah et al. in IEEE Access, vol. 10, pp. 65426–65438 (2022)). This research proposes a deep complex neural network model, namely, advanced semantic segmentation derived from an efficient B0 network for correct identification and detection of brain tumors from MRI images. Image enhancement techniques were employed to improve image quality and training data variability. With the use of enhancement techniques, the size increases. The other DL models included in the comparative analysis are VGG16, InceptionV3, Xception, ResNet50, and InceptionResNetV2.