A Novel 3D InceptionNet Model: Architecture and Applications in Hyperspectral Brain Image Classification

Glioblastoma is an aggressive type of brain cancer with a high mortality rate. Early and accurate glioblastoma detection is crucial for timely and effective treatment. Hyperspectral Imaging (HSI) has emerged as a promising tool for brain tumor detection, as it can capture detailed spectral information that can be leveraged for classification. This paper proposes a novel modified deep learning architecture named 3D InceptionNet to classify hyperspectral brain images. 3D InceptionNet combines the advantages of 3D convolutional layers and the InceptionNet module to extract and learn discriminative features from the 3D hyperspectral data. To demonstrate the efficacy of our proposed approach, we evaluated its performance on an in-vivo hyperspectral brain image dataset using various performance measures.