Characterized Crowd Instance-level Human Parsing (CCIHP) dataset
CCIHP dataset provides pixelwise image annotations for:
- human segmentation,
- semantic attribute segmentation and
- semantic attribute characterization.
Data were annotated with the open-source tool pixano.
Dataset description
-
Images:
The image data are the same as CIHP dataset (see Section Related work) proposed at the LIP (Look Into Person) challenge. They are available at google drive and baidu drive. (Baidu link does not need access right). -
Annotations:
The CCIHP annotations can be found in theTraining
andValidation
sub-folders ofCCIHP_icip2021/dataset/
folder. Annotations consist of:- Human_ids: person instance labels
- Instance_ids: attribute instance labels
- Category_ids: attribute category labels
- Size_ids: size category labels
- Pattern_ids: pattern category labels
- Color_ids: color category labels
Label meaning for semantic attribute/body parts:
- Hat: Hat, helmet, cap, hood, veil, headscarf, part covering the skull and hair of a hood/balaclava, crown…
- Hair
- Glove
- Sunglasses/Glasses: Sunglasses, eyewear, protective glasses…
- UpperClothes: T-shirt, shirt, tank top, sweater under a coat, top of a dress…
- Face Mask: Protective mask, surgical mask, carnival mask, facial part of a balaclava, visor of a helmet…
- Coat: Coat, jacket worn without anything on it, vest with nothing on it, a sweater with nothing on it…
- Socks
- Pants: Pants, shorts, tights, leggings, swimsuit bottoms… (clothing with 2 legs)
- Torso-skin
- Scarf: Scarf, bow tie, tie…
- Skirt: Skirt, kilt, bottom of a dress…
- Face
- Left-arm
- Right-arm
- Left-leg
- Right-leg
- Left-shoe
- Right-shoe
- Bag: Backpack, shoulder bag, fanny pack… (bag carried on oneself)
- Others: Jewelry, tags, bibs, belts, ribbons, pins, head decorations, headphones…
Label meaning for size characterization:
- Short: Small, short, narrow
- Long: Long, large, big
- Undetermined: If the attribute is partially hidden
- Sparse/bald: For hair attribute only
Label meaning for pattern characterization:
- Solid
- Geometrical: Stripes, Checks, Dots…
- Fancy: Flowers, Military…
- Letters: Letters, numbers, symbols…
Label meaning for color characterization:
- Dark: No dominant color, includes black, navy blue
- Medium: No dominant color, including gray
- Light: No dominant color, including white
- Brown
- Red
- Pink
- Yellow
- Orange
- Green
- Blue
- Purple
- Multicolor: When there is a pattern with several colors
Related work
Our work is based on CIHP image dataset from: Ke Gong, Xiaodan Liang, Yicheng Li, Yimin Chen, Ming Yang and Liang Lin, “Instance-level Human Parsing via Part Grouping Network”, ECCV 2018.
Evaluation
To evaluate the predictions given by a Human Parsing with Characteristics model, you can run the python scripts in CCIHP_icip2021/evaluation/
folder.
Requirements
- python==3.6+
- opencv-python
- pillow
Evaluation steps
- Run
generate_characteristic_instance_part_ccihp.py
- Run
eval_test_characteristic_inst_part_ap_ccihp.py
for mean Average Precision based on characterized region (AP^(cr)_(vol)). It evaluates the prediction of characteristic (class & score) relative to each instanced and characterized attribute mask, independently of the attribute class prediction. - Run
metric_ccihp_miou_evaluation.py
for a mIoU performance evaluation of semantic predictions (attribute or characteristics).
License
Data annotations are under Creative Commons Attribution Non Commercial 4.0 license.
Evaluation codes are under MIT license.
Citation
A. Loesch and R. Audigier, “Describe Me If You Can! Characterized Instance-Level Human Parsing,” 2021 IEEE International Conference on Image Processing (ICIP), 2021, pp. 2528-2532, doi: 10.1109/ICIP42928.2021.9506509.
@INPROCEEDINGS{ccihp_dataset_2021,
author={Loesch, Angelique and Audigier, Romaric},
booktitle={2021 IEEE International Conference on Image Processing (ICIP)},
title={Describe Me If You Can! Characterized Instance-Level Human Parsing},
year={2021},
volume={},
number={},
pages={2528-2532},
doi={10.1109/ICIP42928.2021.9506509}},
Contact
If you have any question about this dataset, you can contact us by email at: ccihp-dataset@cea.fr