Abstract — In this paper we study the use of convolutional neural networks (convnets) for the task of pedestrian detection. Despite their recent diverse successes, convnets historically underperform compared to other pedestrian detectors. We deliberately omit explicitly modelling the problem into the network (e.g. parts or occlusion modelling) and show that we can reach competitive performance without bells and whistles. In a wide range of experiments we analyse small and big convnets, their architectural choices, parameters, and the influence of different training data, including pre-training on surrogate tasks.
We present the best convnet detectors on the Caltech and KITTI dataset. On Caltech our convnets reach top performance both for the Caltech1x and Caltech10x training setup. Using additional data at training time our strongest convnet model is competitive even to detectors that use additional data (optical flow) at test time.
Data & Downloads
Proposals used for training and testing
- Alexnet trained from scratch on Caltech10x (warp, ~27.5% MR): caffe net, prototxt
- Alexnet finetuned on Caltech 10x (warp, ~23.3% MR): caffe net, prototxt
- Alexnet finetuned on Caltech 10x (square, ~21.4% MR): caffe net, prototxt
- Alexnet finetuned on Caltech 10x (23.3% MR)
- Alexnet trained from scratch on Caltech10x (27.5% MR)
- Alexnet trained from scratch on Caltech1x (32.4% MR)
- CifarNet trained on Caltech10x (28.4% MR)
- CifarNet trained on Caltech1x (30.7% MR)