MnasFPN: Learning Latency-aware Pyramid Architecture for Object Detection on Mobile Devices

CVPR 2020  ·  Bo Chen, Golnaz Ghiasi, Hanxiao Liu, Tsung-Yi Lin, Dmitry Kalenichenko, Hartwig Adams, Quoc V. Le ·

Despite the blooming success of architecture search for vision tasks in resource-constrained environments, the design of on-device object detection architectures have mostly been manual. The few automated search efforts are either centered around non-mobile-friendly search spaces or not guided by on-device latency... We propose MnasFPN, a mobile-friendly search space for the detection head, and combine it with latency-aware architecture search to produce efficient object detection models. The learned MnasFPN head, when paired with MobileNetV2 body, outperforms MobileNetV3+SSDLite by 1.8 mAP at similar latency on Pixel. It is also both 1.0 mAP more accurate and 10% faster than NAS-FPNLite. Ablation studies show that the majority of the performance gain comes from innovations in the search space. Further explorations reveal an interesting coupling between the search space design and the search algorithm, and that the complexity of MnasFPN search space may be at a local optimum. read more

PDF Abstract CVPR 2020 PDF CVPR 2020 Abstract


Results from the Paper

Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Object Detection COCO test-dev MnasFPN (MobileNetV3) box AP 25.5 # 172
Object Detection COCO test-dev MnasFPN x0.7 (MobileNetV2) box AP 23.8 # 174
Object Detection COCO test-dev MnasFPN (MNASNet-B1) box AP 24.6 # 173
Object Detection COCO test-dev MnasFPN (MobileNetV2) box AP 26.1 # 170