Closed skcskc7 closed 6 years ago
For now, we have person(human)-only models for keypoint detection. For example if you run
python2 tools/infer_simple.py --cfg configs/12_2017_baselines/e2e_keypoint_rcnn_X-101-32x8d-FPN_s1x.yaml \
--output-dir /tmp/detectron-visualizations \
--image-ext jpg \
--wts https://s3-us-west-2.amazonaws.com/detectron/37732318/12_2017_baselines/e2e_keypoint_rcnn_X-101-32x8d-FPN_s1x.yaml.16_55_09.Lx8H5JVu/output/train/keypoints_coco_2014_train:keypoints_coco_2014_valminusminival/generalized_rcnn/model_final.pkl \
demo
You should get output like this:
In the future we will release person-only models that jointly predict detections, keypoints, and segmentation masks.
I want to classify only human class
As a result of running infer_simply.py, i want to a image that contains only human class and human segmentation (does not contain other class)
How can i do this?