The large-scale surveillance videos analysis becomes important as the development of the intelligent city; however, the heavy computational resources necessary for the state-of-the-art deep learning model makes real-time processing hard to be implemented. As the characteristic of high scene similarity generally existing in surveillance videos, we propose an effective compression architecture called dynamic convolution, which can reuse the previous feature maps to reduce the calculation amount; and combine with filter pruning to further speed up the performance. In this paper, we tested the presented method on 45 surveillance videos with various scenes. The experimental results show that the hybrid pruning architecture can reduce up to 80.4% of FLOPs while preserving the precision within 1.34% mAP; furthermore, the method can improve the processing speed up to 2.8 times compared to the traditional Single Shot MultiBox Detection.

随着智能城市的发展，大规模的监控录像分析变得越来越重要。但是，最新的深度学习模型所需的大量计算资源使实时处理难以实现。针对监控视频普遍存在的高场景相似性的特点，我们提出了一种有效的压缩架构，称为动态卷积，可以重用以前的特征图以减少计算量。并与过滤修剪相结合，以进一步提高性能。在本文中，我们在45个具有各种场景的监控视频上测试了该方法。实验结果表明，混合修剪架构可以减少高达80.4％的FLOP，同时将精度保持在1.34％mAP之内。此外，