End-to-end One-shot Human Parsing

Previous human parsing models are limited to parsing humans into pre-defined classes, which is inflexible for practical fashion applications that often have new fashion item classes. In this paper, we define a novel one-shot human parsing (OSHP) task that requires parsing humans into an open set of classes defined by any test example. During training, only base classes are exposed, which only overlap with part of the test-time classes. To address three main challenges in OSHP, i.e., small sizes, testing bias, and similar parts, we devise an End-to-end One-shot human Parsing Network (EOP-Net). Firstly, an end-to-end human parsing framework is proposed to parse the query image into both coarse-grained and fine-grained human classes, which builds a strong embedding network with rich semantic information shared across different granularities, facilitating identifying small-sized human classes. Then, we propose learning momentum-updated prototypes by gradually smoothing the training time static prototypes, which helps stabilize the training and learn robust features. Moreover, we devise a dual metric learning scheme which encourages the network to enhance features' both representational capability and transferability. Therefore, our EOP-Net can learn representative features that can quickly adapt to the novel classes and mitigate the testing bias issue. In addition, we employ a contrastive loss at the prototype level, thereby enforcing the distances among the classes in the fine-grained metric space to discriminate similar parts. We tailor three existing popular human parsing benchmarks to the OSHP task. Experiments on the new benchmarks demonstrate that EOP-Net outperforms representative one-shot segmentation models by large margins, which serves as a strong baseline for further research on this new task. The source code is available at https://github.com/Charleshhy/One-shot-Human-Parsing.