Surpassing Real-World Source Training Data: Random 3D Characters for Generalizable Person Re-Identification
This addresses the data scarcity and privacy issues in person re-identification for surveillance applications, offering a novel synthetic data approach that improves generalization.
The authors tackled the problem of limited generalization in person re-identification due to expensive and privacy-sensitive real-world data by synthesizing a large-scale virtual dataset called RandPerson, which includes 1,801,816 images of 8,000 identities, and showed that models trained on this data generalize better to unseen target domains, outperforming those trained on real datasets like CUHK03, Market-1501, and DukeMTMC-reID.
Person re-identification has seen significant advancement in recent years. However, the ability of learned models to generalize to unknown target domains still remains limited. One possible reason for this is the lack of large-scale and diverse source training data, since manually labeling such a dataset is very expensive and privacy sensitive. To address this, we propose to automatically synthesize a large-scale person re-identification dataset following a set-up similar to real surveillance but with virtual environments, and then use the synthesized person images to train a generalizable person re-identification model. Specifically, we design a method to generate a large number of random UV texture maps and use them to create different 3D clothing models. Then, an automatic code is developed to randomly generate various different 3D characters with diverse clothes, races and attributes. Next, we simulate a number of different virtual environments using Unity3D, with customized camera networks similar to real surveillance systems, and import multiple 3D characters at the same time, with various movements and interactions along different paths through the camera networks. As a result, we obtain a virtual dataset, called RandPerson, with 1,801,816 person images of 8,000 identities. By training person re-identification models on these synthesized person images, we demonstrate, for the first time, that models trained on virtual data can generalize well to unseen target images, surpassing the models trained on various real-world datasets, including CUHK03, Market-1501, DukeMTMC-reID, and almost MSMT17. The RandPerson dataset is available at https://github.com/VideoObjectSearch/RandPerson.