Abstract
The detection ability of multimodal image is not good under low light intensity. In order to improve the target recognition rate of multi-modal images, a multi-modal image target recognition method based on asynchronous deep reinforcement learning is proposed. The edge contour detection model of multi-modal image is established, and the light intensity of multi-modal image is adaptive fusion in the atmosphere scattering environment, and the information enhancement of multi-modal image in low mode is carried out by template matching. In this technique, scene contour feature matching method is used to refine the multimodal image, image features are extracted by fuzzy information tracking method, and significant transmission analysis is performed by brightness component. A multi-modal image target recognition method based on asynchronous deep reinforcement learning is proposed. Experiments show that the image has higher resolution and shorter processing time, which effectively improves the ability of image target recognition.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
REFERENCES
Yu, M. and Zhang, H., HDR imaging based on low-rank matrix completion and total variation constraint, Comput. Eng., 2019, vol. 45, no. 4, pp. 262–266.
Yan, S., Xu, D., Zhang, B., Zhang, H.-J., Yang, Q., and Lin, S., Graph embedding and extensions: a general framework for dimensionality reduction, IEEE Trans. Pattern Anal. Mach. Intell., 2006, vol. 29, no. 1, pp. 40–51. https://doi.org/10.1109/TPAMI.2007.250598
Li, B., Wang, C., and Huang, D-S., Supervised feature extraction based on orthogonal discriminant projection, Neurocomputing, 2009, vol. 73, nos. 1–3, pp. 191–196. https://doi.org/10.1016/j.neucom.2008.09.030
Hou, C., Nie, F., Li, X., Yi, D., and Wu, Y., Joint embedding learning and sparse regression: a framework for unsupervised feature selection, IEEE Trans. Cybern., 2014, vol. 44, no. 6, pp. 793–804. https://doi.org/10.1109/TCYB.2013.2272642
Shang, F., Guo, H., Li, G., Zhang, L., Novel image segmentation method with noise based on one-class SVM, J. Comput. Appl., 2019, vol. 39, no. 3, pp. 874–881. https://doi.org/10.11772/j.issn.1001-9081.2018071494
Li, Z., Study on laser three-dimensional motion imaging technology, Laser Mag., 2018, vol. 39, no. 10, pp. 119–123.
Shan, Y. and Wang, J., Robust object tracking method of adaptive scale and direction, Comput. Eng. Appl., 2018, vol. 54, no. 21, pp. 208–216.
Dai, S., Lü, K., Zhai, R., and Dong, J., Lung segmentation method based on 3D region growing method and improved convex hull algorithm, J. Electron. Inf. Technol., 2016, vol. 38, no. 9, pp. 2358–2364. https://doi.org/10.11999/JEIT151365
Yang, J., Zhao, J., Qiang, Y., and Wang, Q., Lung CT image segmentation combined multi-scale watershed method and region growing method, Comput. Eng. Des., 2014, vol. 35, no. 1, pp. 213–217.
Dai, S., Lu, K., Dong, J., Zhang, Y., and Chen, Y., A novel approach of lung segmentation on chest CT images using graph cuts, Neurocomputing, 2015, vol. 168, pp. 799–807. https://doi.org/10.1016/j.neucom.2015.05.044
Gong, X. and Wang, G., 3D face deformable model based on feature points, J. Software, 2009, vol. 20, no. 3, pp. 724–733.
Li, Xinxin. and Gong, Xun., 3D face modeling and validation in cross-pose face matching, J. Comput. Appl., 2017, vol. 37, no. 1, pp. 262–267. https://doi.org/10.11772/j.issn.1001-9081.2017.01.0262
Gan, J.Y., Zhai, Y.K., Xiang, L., Cao, H., He, G., Zeng, J., and Deng, W., Spatial-temporal texture cascaded feature method for face liveness detection, Pattern Recognit. Artif. Intell., 2019, vol. 32, no. 2, pp. 117–123.
Radenović, F., Tolias, G., and Chum, O., CNN image retrieval learns from Bow: unsupervised fine-tuning with hard examples, Computer Vision – ECCV 2016, Leibe, B., Matas, J., Sebe, N., and Welling, M., Eds., Lecture Notes in Computer Science, vol. 9905, Cham: Springer, 2016, pp. 3–20. https://doi.org/10.1007/978-3-319-46448-0_1
Razavian, A., Sullivan, J., and Carlsson, S., Visual instance retrieval with deep convolutional networks, ITE Trans. Media Technol. Appl., 2016, vol. 4, no. 3, pp. 251–258. https://doi.org/10.3169/mta.4.251
Funding
This work was supported by the Innovation Fund for Production, Study and Research of Chinese Universities of the Science and Technology Development Center of Ministry of Education, Innovation Project for New Generation’s Information Technology (no. 2019ITA03027), the Teaching Research Project of Hubei Provincial Department of Education (no. 2020806), the Teaching Research Project of College of Technology, Hubei Engineering University (no. 2020JY04).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
The authors declare that they have no conflicts of interest.
About this article
Cite this article
Guotao Zhao, Jie Ding Research on Multi-Modal Image Target Recognition Based on Asynchronous Depth Reinforcement Learning. Aut. Control Comp. Sci. 56, 253–260 (2022). https://doi.org/10.3103/S0146411622030105
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.3103/S0146411622030105