An object detection approach with residual feature fusion and second‐order term attention mechanism
Abstract Automatically detecting and locating remote occlusion small objects from the images of complex traffic environments is a valuable and challenging research. Since the boundary box location is not sufficiently accurate and it is difficult to distinguish overlapping and occluded objects, the a...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2024-04-01
|
| Series: | CAAI Transactions on Intelligence Technology |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/cit2.12236 |
| _version_ | 1797201299468451840 |
|---|---|
| author | Cuijin Li Zhong Qu Shengye Wang |
| author_facet | Cuijin Li Zhong Qu Shengye Wang |
| author_sort | Cuijin Li |
| collection | DOAJ |
| description | Abstract Automatically detecting and locating remote occlusion small objects from the images of complex traffic environments is a valuable and challenging research. Since the boundary box location is not sufficiently accurate and it is difficult to distinguish overlapping and occluded objects, the authors propose a network model with a second‐order term attention mechanism and occlusion loss. First, the backbone network is built on CSPDarkNet53. Then a method is designed for the feature extraction network based on an item‐wise attention mechanism, which uses the filtered weighted feature vector to replace the original residual fusion and adds a second‐order term to reduce the information loss in the process of fusion and accelerate the convergence of the model. Finally, an objected occlusion regression loss function is studied to reduce the problems of missed detections caused by dense objects. Sufficient experimental results demonstrate that the authors’ method achieved state‐of‐the‐art performance without reducing the detection speed. The mAP@.5 of the method is 85.8% on the Foggy_cityscapes dataset and the mAP@.5 of the method is 97.8% on the KITTI dataset. |
| first_indexed | 2024-04-24T07:45:20Z |
| format | Article |
| id | doaj.art-1033f408a0894c8090a7115f37821d1e |
| institution | Directory Open Access Journal |
| issn | 2468-2322 |
| language | English |
| last_indexed | 2024-04-24T07:45:20Z |
| publishDate | 2024-04-01 |
| publisher | Wiley |
| record_format | Article |
| series | CAAI Transactions on Intelligence Technology |
| spelling | doaj.art-1033f408a0894c8090a7115f37821d1e2024-04-19T03:11:28ZengWileyCAAI Transactions on Intelligence Technology2468-23222024-04-019241142410.1049/cit2.12236An object detection approach with residual feature fusion and second‐order term attention mechanismCuijin Li0Zhong Qu1Shengye Wang2College of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing ChinaCollege of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing ChinaCollege of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing ChinaAbstract Automatically detecting and locating remote occlusion small objects from the images of complex traffic environments is a valuable and challenging research. Since the boundary box location is not sufficiently accurate and it is difficult to distinguish overlapping and occluded objects, the authors propose a network model with a second‐order term attention mechanism and occlusion loss. First, the backbone network is built on CSPDarkNet53. Then a method is designed for the feature extraction network based on an item‐wise attention mechanism, which uses the filtered weighted feature vector to replace the original residual fusion and adds a second‐order term to reduce the information loss in the process of fusion and accelerate the convergence of the model. Finally, an objected occlusion regression loss function is studied to reduce the problems of missed detections caused by dense objects. Sufficient experimental results demonstrate that the authors’ method achieved state‐of‐the‐art performance without reducing the detection speed. The mAP@.5 of the method is 85.8% on the Foggy_cityscapes dataset and the mAP@.5 of the method is 97.8% on the KITTI dataset.https://doi.org/10.1049/cit2.12236artificial intelligencecomputer visionimage processingmachine learningneural networkobject recognition |
| spellingShingle | Cuijin Li Zhong Qu Shengye Wang An object detection approach with residual feature fusion and second‐order term attention mechanism CAAI Transactions on Intelligence Technology artificial intelligence computer vision image processing machine learning neural network object recognition |
| title | An object detection approach with residual feature fusion and second‐order term attention mechanism |
| title_full | An object detection approach with residual feature fusion and second‐order term attention mechanism |
| title_fullStr | An object detection approach with residual feature fusion and second‐order term attention mechanism |
| title_full_unstemmed | An object detection approach with residual feature fusion and second‐order term attention mechanism |
| title_short | An object detection approach with residual feature fusion and second‐order term attention mechanism |
| title_sort | object detection approach with residual feature fusion and second order term attention mechanism |
| topic | artificial intelligence computer vision image processing machine learning neural network object recognition |
| url | https://doi.org/10.1049/cit2.12236 |
| work_keys_str_mv | AT cuijinli anobjectdetectionapproachwithresidualfeaturefusionandsecondordertermattentionmechanism AT zhongqu anobjectdetectionapproachwithresidualfeaturefusionandsecondordertermattentionmechanism AT shengyewang anobjectdetectionapproachwithresidualfeaturefusionandsecondordertermattentionmechanism AT cuijinli objectdetectionapproachwithresidualfeaturefusionandsecondordertermattentionmechanism AT zhongqu objectdetectionapproachwithresidualfeaturefusionandsecondordertermattentionmechanism AT shengyewang objectdetectionapproachwithresidualfeaturefusionandsecondordertermattentionmechanism |