Abstract
The face is a special target. In order to obtain better detection results, researchers have improved and optimized the target detection algorithm, and proposed many excellent face detection algorithms. This paper improves the target detection algorithm Faster RCNN, uses ResNet50 to extract convolution features, performs multi-scale fusion of feature maps of different convolution layers. At the same time, the anchor boxes generated by the region proposal network are changed from the original 9 to 15 to better adapt to the small-scale face detection scene. On this basis, this paper integrates the SRGAN image super-resolution reconstruction module to perform data enhancement on low-resolution faces and form a detection network that can improve the performance of low-resolution small face detection. Train and test on the Wider Face dataset and conduct network comparison experiments. The final analysis of the experimental data shows that using the image deblurring algorithm of ResNet50 and SRGAN, compared with the original model, the accuracy is 2.06% higher than the original model, especially in low-resolution face detection.
Published Version
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.
