The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.

Dataset Card for AISegment.cn - Matting Human datasets

Dataset Description

Quoting the dataset's github (translated by Apple Translator):

This dataset is currently the largest portrait matting dataset, containing 34,427 images and corresponding matting results. The data set was marked by the high quality of Beijing Play Star Convergence Technology Co. Ltd., and the portrait soft segmentation model trained using this data set has been commercialized.

The original images in the dataset are from Flickr, Baidu, and Taobao. After face detection and area cropping, a half-length portrait of 600*800 was generated. The clip_img directory is a half-length portrait image in the format jpg; the matting directory is the corresponding matting file (convenient to confirm the matting quality), the format is png, you should first extract the alpha map from the png image before training.

Dataset Structure

└── data/
    β”œβ”€β”€ clip_img/
    β”‚   └── {group-id}/
    β”‚       └── clip_{subgroup-id}/
    β”‚           └── {group-id}-{img-id}.jpg
    └── matting/
        └── {group-id}/
            └── matting_{subgroup-id}/
                └── {group-id}-{img-id}.png

The input data/clip_img/1803151818/clip_00000000/1803151818-00000003.jpg matches the label data/matting/1803151818/matting_00000000/1803151818-00000003.png

Licensing Information

See authors Github

Downloads last month
51