Towards Efficient U-Nets: A Coupled and Quantized Approach

In this paper, we propose to couple stacked U-Nets for efficient visual landmark localization. The key idea is to globally reuse features of the same semantic meanings across the stacked U-Nets. The feature reuse makes each U-Net light-weighted. Specially, we propose an order- $K$ coupling design to trim off long-distance shortcuts, together with an iterative refinement, that may slice the model size in half. To further improve the efficiency, we quantize the parameters, intermediate features, and gradients of the coupled U-Nets to low bit-width numbers. We validate our approach in two tasks: human pose estimation and facial landmark localization. The results show that our approach achieves comparable state-of-the-art localization accuracy, but using ~70% fewer parameters, ~30% less inference time, ~98% less model size, and saving ~75% training memory compared with benchmark localizers.