tf-cpn
tf-cpn copied to clipboard
Cascaded Pyramid Network for Multi-Person Pose Estimation (CVPR 2018)
where is ”person_detection_minival411_human553.json“ in COCO dataset,only has the right imgpath. Need I change the code in mptest.py (L190)?
I add $cpn_root/lib to $PYTHONPATH I tried to google, and it seems that I need to run make in $cpn_root/lib I did Still not solved yet
I'd like to apply pre-trained model on my datasets. But validation datasets are required image and bbox(json) file both as well. The loading gt_bbox codes are linked each other and...
Have anyone training this model on yourself datasets? I try to train it with my datasets, the results is below, when i trained about 46 epochs. DONE (t=0.82s). Average Precision...
why? gk15 = (23, 23) gk11 = (17, 17) gk9 = (13, 13) gk7 = (9, 9)
``` global_loss = 0. for i, (global_out, label) in enumerate(zip(global_outs, labels)): global_label = label * tf.to_float(tf.greater(tf.reshape(valids, (-1, 1, 1, cfg.nr_skeleton)), 1.1)) global_loss += tf.reduce_mean(tf.square(global_out - global_label)) / len(labels) global_loss /=...
@chenyilun95 Is there any scripts to detect one single image? I found a scrpt here: https://www.aiuai.cn/aifarm445.html I test it but the result seems to be wrong Anybody tried that?
myubuntu:~/Desktop/tf-cpn/models/COCO.res101.384x288.CPN$ python3 mptest.py -d 0 -m log/model_dump/snapshot_350.ckpt Traceback (most recent call last): File "mptest.py", line 21, in from COCOAllJoints import COCOJoints File "/home/myubuntu/Desktop/tf-cpn/models/COCO.res101.384x288.CPN/../../data/COCO/COCOAllJoints.py", line 12, in from pycocotools.coco import COCO...
Hi, thanks for providing the implementation. For some reason I get somewhat strange results when plotting joints. It just draws the perfect straight line through the 17 keypoints (seems to...