All Times ET
Keywords: Brain tumor, Diffusion imaging, Neurocognitive scores, 3D convolutional auto-encoder, Deep learning, Clustering.
The Clustering of pediatric patients based on their unlabeled brain magnetic resonance (MR) images reveal the latent features and group structures. It is an unbiased approach to finding image-based markers and risk-stratification for neurocognitive functions. Analyzing large sets of MR images is time-intensive and challenging due to the MR images' high dimensionality and complexity. The extraction of low-dimension informative feature space instead of high-dimension MR image is an effective and fast solution to cluster images accurately. The ability to reduce the size of large 3D MR images and extract the informative features and information from them is critically important in neuroscience research. This paper developed an automated tool based on a deep learning technique and a 3D convolutional auto-encoder model to analyze large datasets of MR images and cluster them into great clusters. We reduced the size of original MRI data by 256 times and reconstructed them precisely using the reduced size of data. Our long-term research goal is to develop a feature extraction model using deep learning to automatically and accurately identify brain regions associated with neurocognitive outcomes. We clustered the MRI data into several clusters based on these reduced features and clustering algorithms like K-means and Sparse K-means. After that, we considered some neurocognitive data and found a remarkable correlation between image and neurocognitive datasets.