Action Detection Based on 3D Convolution Neural Network with Channel Attention Mechanism

Brain MR images segmentation using 3D CNN with features recalibration mechanism for segmented CT generation

“…Hu et al [19] were the first who have introduced the features recalibration approach using channel excitation block for ILSVRC 2017 classification competition. Since then, this mechanism has been widely and successfully applied to solve different applications such as breast density categories classification [23], action detection [24], and sea ice images classification [25]. The spatial squeeze and channel excitation also has been applied for various medical images segmentation applications such as brain tumor segmentation [26], prostate zonal segmentation [27] and pancreas segmentation using CT images [28].…”