The current infrared image pedestrian detectors have problems with high rates of false positives and false negatives. To solve these problems, we proposed an improved anchor-free fully convolutional one-stage object detection (FCOS) algorithm. Firstly, we introduced the channel attention module squeeze excitation (SE)-Block in the FCOS backbone network, which was used to learn how to model the relative importance between different feature channels, and to achieve the weight recalibration of the features extracted from the convolution neural network, and improve the weight values that are more important for pedestrian target detection. Secondly, soft non-maximum suppression (Soft-NMS) replaced the conventional NMS within the algorithm’s post-processing phase, which was used to reduce the probability of missed detection for occluded pedestrians. The experimental results show that our improved FCOS algorithm improves the average precision (AP) by 6.71% on the original dataset and 7.97% on the augmented KAIST pedestrian dataset compared with the original FCOS algorithm. Our improvements effectively meet the real-time requirements and there is no significant decrease in speed compared with the original FCOS algorithm, and decreased the false positives and false negatives for infrared image pedestrian detection.
| [1] |
Huo L L, Wang Y, Liu Tet al. . Overview of pedestrian detection based on infrared image. 41st Chinese Control Conference (CCC) Information Fusion, July 25–27, 2022, Hefei, China. 2022, New York, IEEE63576362[C]
|
| [2] |
Manssor S A F, Sun S Y, Adbalmajed Met al. . Real-time human detection in thermal infrared imaging at night using enhanced Tiny-yolov3 network. Journal of real-time image processing. 2022, 19(2): 261-274. J]
|
| [3] |
Park S J, Choi D H, Kim J Uet al. . Robust thermal infrared pedestrian detection by associating visible pedestrian knowledge. 2022 IEEE International Conference on Acoustics, Speech and Signal Processing, May 23–27, 2022, Singapore. 2022, New York, IEEE44684472[C]
|
| [4] |
Lee W Y, Jovanov L, Philips W. Multi-view target transformation for pedestrian detection. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, January 2–7, 2023, Waikoloa, USA. 2023, New York, IEEE9099[C]
|
| [5] |
Ren S Q, He K M, Girshick Ret al. . Faster R-CNN: towards real-time object detection with region proposal networks. IEEE transactions on pattern analysis and machine intelligence. 2016, 39(6): 1137-1149. J]
|
| [6] |
Liu W, Anguelov D, Erhan Det al. . SSD: single shot multi-box detector. 14th European Conference on Computer Vision, October 11–14, 2016, Amsterdam, The Netherlands. 2016, Berlin, Springer2137[C]
|
| [7] |
Redmon J, Farhadi A. YOLO9000: better, faster, stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, July 21–26, 2017, Honolulu, USA. 2017, New York, IEEE72637271[C]
|
| [8] |
Zhao L Q, Li S Y. Object detection algorithm based on improved YOLOv3. Electronics. 2020, 9(3): 537. J]
|
| [9] |
Terven J, Córdova D M, Romero J A. A comprehensive review of YOLO architectures in computer vision: from YOLOv1 to YOLOv8 and YOLO-nas. Machine learning and knowledge extraction. 2023, 541680-1716. J]
|
| [10] |
Tian Z, Shen C H, Chen Het al. . FCOS: fully convolutional one-stage object detection. 2019 IEEE/CVF International Conference on Computer Vision, October 27–November 2, 2019, Seoul, Korea. 2019, New York, IEEE96269635[C]
|
| [11] |
Zhang Y G, Zhai B, Wang Get al. . Pedestrian detection method based on two-stage fusion of visible light image and thermal infrared image. Electronics. 2023, 12(14): 3171. J]
|
| [12] |
Liu Z Y, Dai C Y, Li X. Pedestrian detection method in infrared image based on improved YOLOv7. 2023 IEEE 3rd International Conference on Information Technology, Big Data and Artificial Intelligence, May 26–28, 2023, Chongqing, China. 2023, New York, IEEE946-954[C]
|
| [13] |
Zhang Y H, Ji K, He Z Fet al. . Attention-guided multi-scale infrared real-time detection of pedestrian and vehicle. Infrared and laser engineering. 2024, 53(05): 237-247[J]
|
| [14] |
Zhou L, Gao S, Wang S Met al. . IPD-Net: infrared pedestrian detection network via adaptive feature extraction and coordinate information fusion. Sensors. 2022, 22(22): 8966. J]
|
| [15] |
Yao S B, Zhu Q Y, Zhang Tet al. . Infrared image small-target detection based on improved FCOS and spatial-temporal features. Electronics. 2022, 116933. J]
|
| [16] |
Qi X H, Zhi M. A review of attention mechanisms in computer vision. 2023 8th International Conference on Image, Vision and Computing, July 27–29, 2023, Dalian, China. 2023, New York, IEEE577-583[C]
|
| [17] |
Guo M H, Xu T X, Liu J Jet al. . Attention mechanisms in computer vision: a survey. Computational visual media. 2022, 8(3): 331-368. J]
|
| [18] |
Hu J, Shen L, Sun G. Squeeze-and-excitation networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, June 18–22, 2018, Salt Lake City, USA. 2018, New York, IEEE71327141[C]
|
| [19] |
Chen F X, Zhang L X, Kang S Yet al. . Soft-NMS-enabled YOLOv5 with SIOU for small water surface floater detection in UAV-captured images. Sustainability. 2023, 15(14): 10751. J]
|
| [20] |
Shin U, Park J, Kweon I S. Deep depth estimation from thermal image. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 18–22, Vancouver, Canada. 2023, New York, IEEE1043-1053[C]
|
| [21] |
Bai X F, Wu K J, Bai C S. Proposition of remote sensing image object detection algorithm based on EYOLOv3i. 2023 6th International Conference on Data Science and Information Technology, July 28–30, 2023, Shanghai, China. 2023, New York, IEEE243248[C]
|
RIGHTS & PERMISSIONS
Tianjin University of Technology
PDF
