Abstract
Recently, building an efficient and robust model for object detection has attracted the attention of the vision community. Although binary networks have a fast inference speed, they cannot be used directly on mobile devices such as unmanned aerial vehicles (UAVs) because of their low detection accuracy. Different from improving the detection accuracy of a binary network by adjusting the network structure or adjusting the update gradient, we propose an improved binary neural network based on the block scaling factor XNOR (BSF-XNOR) convolutional layer. In addition, we propose a two-level densely connected network structure, which further enhances the network layer's feature representation capabilities. Experiments using the TensorFlow framework prove the effectiveness of our algorithm in improving accuracy. Compared with the original standard XNOR network, the mean average precision (mAP) detected by our algorithm on the PASCAL VOC dataset was improved. The experimental results on the VisDrone2019 UAV dataset confirm that our method achieves a better balance between speed and accuracy than previous methods. Our algorithm aims to guide and deploy high-precision binary networks on the embedded device and solves the problem of low-precision binary networks.
Highlights
Unmanned aerial vehicles (UAVs) have played an important role in both local and national defense fields, and they have been applied to intelligent transportation, geological exploration, military guidance, and aviation visual navigation
EXPERIMENTAL RESULTS ON THE VisDrone2019 DATASET To further show the efficiency of our detection method of the BSF-XNOR binary algorithm, we show the comparisons between our method and existing state-of-the-art methods on the VisDrone UAV dataset, which consists of 10 classes
The algorithm has relatively high accuracy and a 2.4 M parameters, which makes it possible for the algorithm to run in real time on the CPU and other embedded devices such as field-programmable gate arrays (FPGAs)
Summary
Unmanned aerial vehicles (UAVs) have played an important role in both local and national defense fields, and they have been applied to intelligent transportation, geological exploration, military guidance, and aviation visual navigation. The algorithm performs well on embedded devices, the weight and feature activation values are compressed to eight bits, the running speed is increased four- to eightfold, and the object detection accuracy shows almost no drop. The DoReFa algorithm quantizes the gradient and improves the training speed and accuracy [30], achieving an accuracy of 45% on the ImageNet dataset classification task [31] This result is caused by a massive loss of feature map information in the process of binarization. [32] proposed the densely connected algorithm; the algorithm achieved a precision of 58.6% and won first place on the classification task of the ImageNet dataset These papers suggest that increasing representation capabilities can greatly improve the classification or detection accuracy.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
