Study on the Lightweighting Strategy of Target Detection Model with Deep Learning
Aiming at the high miss detection and false detection rate of traditional SSD (single shot multibox detector) target detection algorithm in target detection, this paper proposes a lightweight detection algorithm for deep learning target detection model in order to improve the detection accuracy. Fir...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2022-01-01
|
| Series: | Advances in Multimedia |
| Online Access: | http://dx.doi.org/10.1155/2022/7234888 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849306045845864448 |
|---|---|
| author | Junli Hu |
| author_facet | Junli Hu |
| author_sort | Junli Hu |
| collection | DOAJ |
| description | Aiming at the high miss detection and false detection rate of traditional SSD (single shot multibox detector) target detection algorithm in target detection, this paper proposes a lightweight detection algorithm for deep learning target detection model in order to improve the detection accuracy. Firstly, the real-time and efficient target detection backbone network VoVNet is used to replace the feature extraction network VGG16. The residual structure is integrated to solve the problem of VoVNet network degradation to improve network performance. Secondly, self-attention mechanism is introduced to capture multiscale local and global information to obtain richer image semantic features. According to the characteristics of the target sample size, the anchor frame is designed by using a priori information. In the network training, the anchor frame by selection enhancement based on IoU optimization is used to fully train the target information and strengthen the reading ability of the network to small-scale targets. Experiments on the public data set AI-TOD to show that the target detection lightweight model of deep learning has stronger detection ability and higher average detection accuracy than other algorithms, which proves the applicability and effectiveness of this algorithm. |
| format | Article |
| id | doaj-art-33c5927b8a2a4b58b707e23af9254da9 |
| institution | Kabale University |
| issn | 1687-5699 |
| language | English |
| publishDate | 2022-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Advances in Multimedia |
| spelling | doaj-art-33c5927b8a2a4b58b707e23af9254da92025-08-20T03:55:12ZengWileyAdvances in Multimedia1687-56992022-01-01202210.1155/2022/7234888Study on the Lightweighting Strategy of Target Detection Model with Deep LearningJunli Hu0Henan Industry and Trade Vocational CollegeAiming at the high miss detection and false detection rate of traditional SSD (single shot multibox detector) target detection algorithm in target detection, this paper proposes a lightweight detection algorithm for deep learning target detection model in order to improve the detection accuracy. Firstly, the real-time and efficient target detection backbone network VoVNet is used to replace the feature extraction network VGG16. The residual structure is integrated to solve the problem of VoVNet network degradation to improve network performance. Secondly, self-attention mechanism is introduced to capture multiscale local and global information to obtain richer image semantic features. According to the characteristics of the target sample size, the anchor frame is designed by using a priori information. In the network training, the anchor frame by selection enhancement based on IoU optimization is used to fully train the target information and strengthen the reading ability of the network to small-scale targets. Experiments on the public data set AI-TOD to show that the target detection lightweight model of deep learning has stronger detection ability and higher average detection accuracy than other algorithms, which proves the applicability and effectiveness of this algorithm.http://dx.doi.org/10.1155/2022/7234888 |
| spellingShingle | Junli Hu Study on the Lightweighting Strategy of Target Detection Model with Deep Learning Advances in Multimedia |
| title | Study on the Lightweighting Strategy of Target Detection Model with Deep Learning |
| title_full | Study on the Lightweighting Strategy of Target Detection Model with Deep Learning |
| title_fullStr | Study on the Lightweighting Strategy of Target Detection Model with Deep Learning |
| title_full_unstemmed | Study on the Lightweighting Strategy of Target Detection Model with Deep Learning |
| title_short | Study on the Lightweighting Strategy of Target Detection Model with Deep Learning |
| title_sort | study on the lightweighting strategy of target detection model with deep learning |
| url | http://dx.doi.org/10.1155/2022/7234888 |
| work_keys_str_mv | AT junlihu studyonthelightweightingstrategyoftargetdetectionmodelwithdeeplearning |