A Hierarchical Deep Architecture and Mini-Batch Selection Method For Joint Traffic Sign and Light Detection

  • 2018-09-13 17:20:44
  • Alex D. Pon, Oles Andrienko, Ali Harakeh, Steven L. Waslander
  • 0


Traffic light and sign detectors on autonomous cars are integral for roadscene perception. The literature is abundant with deep learning networks thatdetect either lights or signs, not both, which makes them unsuitable forreal-life deployment due to the limited graphics processing unit (GPU) memoryand power available on embedded systems. The root cause of this issue is thatno public dataset contains both traffic light and sign labels, which leads todifficulties in developing a joint detection framework. We present a deephierarchical architecture in conjunction with a mini-batch proposal selectionmechanism that allows a network to detect both traffic lights and signs fromtraining on separate traffic light and sign datasets. Our method solves theoverlapping issue where instances from one dataset are not labelled in theother dataset. We are the first to present a network that performs jointdetection on traffic lights and signs. We measure our network on theTsinghua-Tencent 100K benchmark for traffic sign detection and the Bosch SmallTraffic Lights benchmark for traffic light detection and show it outperformsthe existing Bosch Small Traffic light state-of-the-art method. We focus onautonomous car deployment and show our network is more suitable than othersbecause of its low memory footprint and real-time image processing time.Qualitative results can be viewed at https://youtu.be/_YmogPzBXOw


Introduction (beta)



Conclusion (beta)