mscoco_keypoint / README.md
WeiChow's picture
Update README.md
0c025d3 verified
metadata
license: apache-2.0
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: val
        path: data/val-*
dataset_info:
  features:
    - name: id
      dtype: string
    - name: caption
      dtype: string
    - name: image
      dtype: image
    - name: cond
      dtype: image
  splits:
    - name: val
      num_bytes: 166271415.25
      num_examples: 2603
    - name: train
      num_bytes: 4007778277
      num_examples: 62312
  download_size: 4097346656
  dataset_size: 4174049692.25

We employed DWPose to detect keypoints in the MSCOCO dataset, subsequently filtering out images where no keypoints were successfully detected.

# detect condition
keypoint_image = annotator(image, width=image.width, height=image.height)
if keypoint_image.dtype != np.uint8:
    keypoint_image = (keypoint_image * 255).clip(0, 255).astype(np.uint8)
keypoint_image = Image.fromarray(keypoint_image)

is_all_black = np.all(np.array(keypoint_image) == 0)