Fig. 1From: Automatic localization and identification of mitochondria in cellular electron cryo-tomography using faster-RCNNFlowchart of our Faster-RCNN model. The denoised input image is fed into Conv layers to generate the feature map. Then, region proposal network proposes potential regions that contain object of interest. The proposal regions are passed to 1) classifier for classification, 2) regressor for refine the bounding box locationBack to article page