Using CNN with Bayesian optimization to identify cerebral micro-bleeds

摘要: This article studies the problem of detecting cerebral micro-bleeds (CMBs) using a convolutional neural network (CNN). Cerebral are increasingly recognized neuroimaging findings, occurring with cerebrovascular diseases, dementia, and normal aging. Naturally enough, it becomes necessary to detect CMBs in early stages life. The focus this is infuse new techniques like Bayesian optimization find optimum set hyper-parameters efficiently, making even simplest CNN architectures perform well on problem. Experimentally, we observe our (five layers, i.e., two convolution, pooling, one fully connected) achieves accuracy = 98.97%, sensitivity 99.66%, specificity 98.14%, precision 98.54% test (hold-out validation) when calculated over an average ten runs. proposed model outperformed state-of-the-art methods.