Details

Autor(en) / Beteiligte

• Rastogi, Deependra
• Johri, Prashant
• Tiwari, Varun
• Singh, Ravendra
• Bhadoria, Vikas Singh
• Shouran, Mokhtar
• Ohene-Akoto, Ing. Justice
• Arunprasad, G
• Ambikapathy, A

Titel

Brain tumor segmentation and enhancing prediction using unet deep learning model with additive skip connection

Ist Teil von

• AIP Conference Proceedings, 2024, Vol.2816 (1)

Ort / Verlag

Melville: American Institute of Physics

Erscheinungsjahr

2024

Beschreibungen/Notizen

• Using multimodal MRI images, the brain tumor segmentation assignment aims to differentiate tumor into the entire tumor (WT), enhancing tumor (ET), and tumor center (TC) classes. For therapeutic decision-making, quantitative examination of brain tumors is important. Though manual segmentation is repetitive, arbitrary, and time-consuming, automatic segmentation approaches face significant challenges. Convolutional neural networks (CNNs) have shown promise in brain tumor segmentation due to their powerful learning capacity. This article suggests an UNet model with addictive skip connection for brain tumor segmentation and enhancing the prediction. The presented model successfully segmented brain tumors and predicted the result in enhancing tumor and real enhancing tumor. Experiment with BRATS2020 dataset, we found that Loss (for Training: .0225, Testing: .0231 and Validation: .0225), Dice Coefficient ( for Training : .9774, Testing : .9768 and Validation : .9773) and Accuracy ( for Training : .9774 Testing : .9768 and Validation : .9773).