sess_input = [model.final_boxes, model.final_labels,
model.final_probs]
final_boxes, final_labels, final_probs = sess.run(
sess_input, feed_dict=feed_dict)
if args.out_dir is None:
cur_frame += 1
continue
# ---------------- get the json outputs for object detection
# scale back the box to original image size
final_boxes = final_boxes / scale
if args.add_mask:
final_masks = [fill_full_mask(box, mask, im.shape[:2])
for box, mask in zip(final_boxes, final_masks)]
# save as json
pred = []
for j, (box, prob, label) in enumerate(zip(
final_boxes, final_probs, final_labels)):
box[2] -= box[0]
box[3] -= box[1] # produce x,y,w,h output
cat_id = int(label)
cat_name = targetid2class[cat_id]
# encode mask
rle = None
final_labels[i] = coco_obj_class_to_id[
args.partial_classes[final_labels[i] - 1]]
else:
# 1-90 to 1-80
for i in range(len(final_labels)):
final_labels[i] = \
coco_obj_class_to_id[coco_id_mapping[final_labels[i]]]
# ---------------- get the json outputs for object detection
# scale back the box to original image size
final_boxes = final_boxes / scale
if args.add_mask:
final_masks = [
fill_full_mask(box, mask, im.shape[:2])
for box, mask in zip(final_boxes, final_masks)
]
# save as json
pred = []
for j, (box, prob, label) in enumerate(
zip(final_boxes, final_probs, final_labels)):
box[2] -= box[0]
box[3] -= box[1] # produce x,y,w,h output
cat_id = int(label)
cat_name = targetid2class[cat_id]
# encode mask
if args.add_mask:
sess_input = [model.final_boxes, model.final_labels, model.final_probs, model.final_masks]
final_boxes, final_labels, final_probs, final_masks = sess.run(sess_input, feed_dict=feed_dict)
else:
sess_input = [model.final_boxes, model.final_labels, model.final_probs]
final_boxes, final_labels, final_probs = sess.run(sess_input, feed_dict=feed_dict)
if args.out_dir is None:
cur_frame += 1
continue
# scale back the box to original image size
final_boxes = final_boxes / scale
if args.add_mask:
final_masks = [fill_full_mask(box, mask, im.shape[:2]) for box, mask in zip(final_boxes, final_masks)]
# save as json
pred = []
for j, (box, prob, label) in enumerate(zip(final_boxes, final_probs, final_labels)):
box[2] -= box[0]
box[3] -= box[1] # produce x,y,w,h output
cat_id = label
cat_name = targetid2class[cat_id]
# encode mask
rle = None
if args.add_mask: