Learning Generalisable Omni-Scale Representations for Person Re-Identification

15 Oct 2019  ·  Kaiyang Zhou, Yongxin Yang, Andrea Cavallaro, Tao Xiang ·

An effective person re-identification (re-ID) model should learn feature representations that are both discriminative, for distinguishing similar-looking people, and generalisable, for deployment across datasets without any adaptation. In this paper, we develop novel CNN architectures to address both challenges. First, we present a re-ID CNN termed omni-scale network (OSNet) to learn features that not only capture different spatial scales but also encapsulate a synergistic combination of multiple scales, namely omni-scale features. The basic building block consists of multiple convolutional streams, each detecting features at a certain scale. For omni-scale feature learning, a unified aggregation gate is introduced to dynamically fuse multi-scale features with channel-wise weights. OSNet is lightweight as its building blocks comprise factorised convolutions. Second, to improve generalisable feature learning, we introduce instance normalisation (IN) layers into OSNet to cope with cross-dataset discrepancies. Further, to determine the optimal placements of these IN layers in the architecture, we formulate an efficient differentiable architecture search algorithm. Extensive experiments show that, in the conventional same-dataset setting, OSNet achieves state-of-the-art performance, despite being much smaller than existing re-ID models. In the more challenging yet practical cross-dataset setting, OSNet beats most recent unsupervised domain adaptation methods without using any target data. Our code and models are released at \texttt{https://github.com/KaiyangZhou/deep-person-reid}.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Unsupervised Person Re-Identification DukeMTMC-reID->Market-1501 OSNet-AIN mAP 30.6 # 5
Rank-1 61 # 4
Rank-10 82.5 # 2
Rank-5 77 # 3
Unsupervised Person Re-Identification MSMT17->DukeMTMC-reID OSNet-AIN Rank-1 70.1 # 1
Rank-10 88.6 # 1
Rank-5 84.1 # 1
mAP 43.3 # 1
Unsupervised Person Re-Identification MSMT17->Market-1501 OSNet-AIN Rank-1 71.1 # 1
Rank-10 86.4 # 1
Rank-5 83.3 # 1
mAP 52.7 # 1


No methods listed for this paper. Add relevant methods here