arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1804.06332cs.CV
收藏

通过知识转移训练二进制权重物体检测器用于自动驾驶

Training a Binary Weight Object Detector by Knowledge Transfer for Autonomous Driving

Computer Vision Center(计算机视觉中心);Autonomous University of Barcelona(巴塞罗那自治大学)

Jiaolong Xu, Peng Wang, Heng Yang, Antonio M. López

摘要:

自动驾驶对模型体积小、能效高有着苛刻的要求,为了使嵌入式系统能够实现实时的车载目标检测。最近的基于深度卷积神经网络的对象检测器已经达到了最先进的精度。然而,这样的模型是用许多参数训练的,并且它们的高计算成本和大存储禁止部署到存储器和计算资源有限的系统。低精度神经网络是用于减少计算需求和存储器占用的流行技术。其中,二进制权值神经网络(BWN)是将浮点数量化为$1$位的极端情况。BWN难以训练,并且由于极低比特表示而遭受精度降低。为了解决这个问题,我们提出了一种知识转移(KT)方法来帮助使用全精度教师网络训练BWN。我们构建了基于DarkNet和MobileNet的二进制权重YOLO-v2检测器,并在KITTI基准上进行了汽车、行人和骑自行车者检测实验。实验结果表明,该方法保持了较高的检测精度,同时将DarkNet-YOLO的模型大小从257 MB减少到8.8 MB,MobileNet-YOLO从193 MB减少到7.9 MB。

英文摘要:

Autonomous driving has harsh requirements of small model size and energy efficiency, in order to enable the embedded system to achieve real-time on-board object detection. Recent deep convolutional neural network based object detectors have achieved state-of-the-art accuracy. However, such models are trained with numerous parameters and their high computational costs and large storage prohibit the deployment to memory and computation resource limited systems. Low-precision neural networks are popular techniques for reducing the computation requirements and memory footprint. Among them, binary weight neural network (BWN) is the extreme case which quantizes the float-point into just $1$ bit. BWNs are difficult to train and suffer from accuracy deprecation due to the extreme low-bit representation. To address this problem, we propose a knowledge transfer (KT) method to aid the training of BWN using a full-precision teacher network. We built DarkNet- and MobileNet-based binary weight YOLO-v2 detectors and conduct experiments on KITTI benchmark for car, pedestrian and cyclist detection. The experimental results show that the proposed method maintains high detection accuracy while reducing the model size of DarkNet-YOLO from 257 MB to 8.8 MB and MobileNet-YOLO from 193 MB to 7.9 MB.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑