I downloaded the tensorflow object detection mask rcnn model and make inference successfully.
The model has 5 outputs, they are num_detections, detection_classes, detection_scores, detection_boxes, detection_masks.
The num_detections, detection_classes, detection_scores, detection_boxes work very well and I got all output result. But detection_masks.dtypr = array(num_detections, 15, 15), the number in it is float and within (0, 1). I don't know hot to use this (15,15) data to draw a mask picture of an object. PLEASE HELP!
The 15x15 box is a grid of the predicted box. So use your box[0] to plot mask[0] using the dimensions of the box. Hope this helps even though it was asked a month ago 👍
I downloaded the tensorflow object detection mask rcnn model and make inference successfully. The model has 5 outputs, they are num_detections, detection_classes, detection_scores, detection_boxes, detection_masks. The num_detections, detection_classes, detection_scores, detection_boxes work very well and I got all output result. But detection_masks.dtypr = array(num_detections, 15, 15), the number in it is float and within (0, 1). I don't know hot to use this (15,15) data to draw a mask picture of an object. PLEASE HELP!