Our work allows computer to recognize clothings in images. This method is based on FCN architecture. The detail of our method and dataset is in https://arxiv.org/abs/1703.01386.
This live model is trained to recognize 20 clothing categories, plus face and skin based on CFPD dataset.
Input: preferably full body frontal image in 2:3 ratio containing only 1 person.
less than 2MB; only '.png', '.bmp', '.jpg', '.jpe', '.jpeg', '.gif' allowed.
Our model took 0.243 seconds.