Multiclass objects detection algorithm using DarkNet-53 and DenseNet for intelligent vehicles
Abstract Intelligent vehicles should not only be able to detect various obstacles, but also identify their categories so as to take an appropriate protection and intervention. However, the scenarios of object detection are usually complex and changeable, so how to balance the relationship between ac...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2023-08-01
|
| Series: | EURASIP Journal on Advances in Signal Processing |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13634-023-01045-8 |
| _version_ | 1797752547186835456 |
|---|---|
| author | Lina Yang Gang Chen Wenyan Ci |
| author_facet | Lina Yang Gang Chen Wenyan Ci |
| author_sort | Lina Yang |
| collection | DOAJ |
| description | Abstract Intelligent vehicles should not only be able to detect various obstacles, but also identify their categories so as to take an appropriate protection and intervention. However, the scenarios of object detection are usually complex and changeable, so how to balance the relationship between accuracy and speed is a difficult task of object detection. This paper proposes a multi-object detection algorithm using DarkNet-53 and dense convolution network (DenseNet) to further ensure maximum information flow between layers. Three 8-layer dense blocks are used to replace the last three downsampling layers in DarkNet-53 structure, so that the network can make full use of multi-layer convolution features before prediction. The loss function of coordinate prediction error in YOLOv3 is further improved to improve the detection accuracy. Extensive experiments are conducted on the public KITTI and Pascal VOC datasets, and the results demonstrate that the proposed algorithm has better robustness, and the network model is more suitable for the traffic scene in the real driving environment and has better adaptability to the objects with long distance, small size and partial occlusion. |
| first_indexed | 2024-03-12T17:05:58Z |
| format | Article |
| id | doaj.art-58e4cb4e13b2446da40099dcf24f4469 |
| institution | Directory Open Access Journal |
| issn | 1687-6180 |
| language | English |
| last_indexed | 2024-03-12T17:05:58Z |
| publishDate | 2023-08-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | EURASIP Journal on Advances in Signal Processing |
| spelling | doaj.art-58e4cb4e13b2446da40099dcf24f44692023-08-06T11:27:29ZengSpringerOpenEURASIP Journal on Advances in Signal Processing1687-61802023-08-012023111810.1186/s13634-023-01045-8Multiclass objects detection algorithm using DarkNet-53 and DenseNet for intelligent vehiclesLina Yang0Gang Chen1Wenyan Ci2College of Information Science and Engineering, Jiaxing UniversityCollege of Information Science and Engineering, Jiaxing UniversityCollege of Engineering, Huzhou UniversityAbstract Intelligent vehicles should not only be able to detect various obstacles, but also identify their categories so as to take an appropriate protection and intervention. However, the scenarios of object detection are usually complex and changeable, so how to balance the relationship between accuracy and speed is a difficult task of object detection. This paper proposes a multi-object detection algorithm using DarkNet-53 and dense convolution network (DenseNet) to further ensure maximum information flow between layers. Three 8-layer dense blocks are used to replace the last three downsampling layers in DarkNet-53 structure, so that the network can make full use of multi-layer convolution features before prediction. The loss function of coordinate prediction error in YOLOv3 is further improved to improve the detection accuracy. Extensive experiments are conducted on the public KITTI and Pascal VOC datasets, and the results demonstrate that the proposed algorithm has better robustness, and the network model is more suitable for the traffic scene in the real driving environment and has better adaptability to the objects with long distance, small size and partial occlusion.https://doi.org/10.1186/s13634-023-01045-8Multiclass objects detectionDarkNet-53DenseNetDownsampling layersLoss function |
| spellingShingle | Lina Yang Gang Chen Wenyan Ci Multiclass objects detection algorithm using DarkNet-53 and DenseNet for intelligent vehicles EURASIP Journal on Advances in Signal Processing Multiclass objects detection DarkNet-53 DenseNet Downsampling layers Loss function |
| title | Multiclass objects detection algorithm using DarkNet-53 and DenseNet for intelligent vehicles |
| title_full | Multiclass objects detection algorithm using DarkNet-53 and DenseNet for intelligent vehicles |
| title_fullStr | Multiclass objects detection algorithm using DarkNet-53 and DenseNet for intelligent vehicles |
| title_full_unstemmed | Multiclass objects detection algorithm using DarkNet-53 and DenseNet for intelligent vehicles |
| title_short | Multiclass objects detection algorithm using DarkNet-53 and DenseNet for intelligent vehicles |
| title_sort | multiclass objects detection algorithm using darknet 53 and densenet for intelligent vehicles |
| topic | Multiclass objects detection DarkNet-53 DenseNet Downsampling layers Loss function |
| url | https://doi.org/10.1186/s13634-023-01045-8 |
| work_keys_str_mv | AT linayang multiclassobjectsdetectionalgorithmusingdarknet53anddensenetforintelligentvehicles AT gangchen multiclassobjectsdetectionalgorithmusingdarknet53anddensenetforintelligentvehicles AT wenyanci multiclassobjectsdetectionalgorithmusingdarknet53anddensenetforintelligentvehicles |