Patch-based vehicle logo detection with patch intensity and weight matrix

Abstract

A patch-based method for detecting vehicle logos using prior knowledge is proposed. By representing the coarse region of the logo with the weight matrix of patch intensity and position, the proposed method is robust to bad and complex environmental conditions. The bounding-box of the logo is extracted by a thershloding approach. Experimental results show that 93.58% location accuracy is achieved with 1100 images under various environmental conditions, indicating that the proposed method is effective and suitable for the location of vehicle logo in practical applications.

Keywords

vehicle logo detection / prior knowledge / gradient extraction / patch intensity / weight matrix / background removing

References

[1] ZhouK, VaradarajanK M, VinczeM, LiuF-qiang. Hybridization of appearance and symmetry for vehicle-logo localization [C]. 15th International IEEE Intelligent Conference on Transportation Systems, 2012Anchorage, AKIEEE1396-1401

[2] YangW-w, YangG-w, ZhengX, ZhangL-bo. An improved vehicle-logo localization algorithm based on texture analysis [C]. International Conference on Computer Science and Information Processing (CSIP), 2012Xi’an, ChinaIEEE648-651

[3] LiuY, LiS-tao. A vehicle-logo location approach based on edge detection and projection [C]. IEEE International Conference on Vehicular Electronics and Safety (ICVES), 2011Beijing, ChinaIEEE165-168

[4] LiW-j, LiLing. A novel approach for vehicle-logo location based on edge detection and morphological filter [C]. Second International Symposium on Electronic Commerce and Security, 2009Nanchang, ChinaIEEE343-345

[5] WangY-q, LiuZ-f, XiaoFei. A fast coarse-to-fine vehicle logo detection and recognition method [C]. IEEE International Conference on Robotics and Biomimetics, 2007Sanya, ChinaIEEE691-696

[6] MaoS-a, YeM, LiX, PangF, ZhouJ-lei. Rapid vehicle logo region detection based on information theory [J]. Computers & Electrical Engineering, 2013, 39: 863-872

[7] LlorcaD F, ArroyoR, SoteloM A. Vehicle logo recognition in traffic images using HOG features and SVM [C]. 16th International IEEE Conference on Intelligent Transportation Systems, 2013The HagueIEEE2229-2234

[8] LuF, LiuY, ZhangR-hui. An improved HOG-based vehicle logo location and recognition method [J]. Study on Optical Communications, 2012, 5: 26-29

[9] YangH-h, ZhaiL, LiuZ-b, LuoY-c, WangY, LiH-y, GuanMing. An efficient method for vehicle model identification via logo recognition [C]. 5th International Conference on Computational and Information Sciences (ICCIS), 2013Shiyang, ChinaIEEE1080-1083

[10] ThubsaengW, KawewongA, PatanukhomK. Vehicle logo detection using convolutional neural network and pyramid of histogram of oriented gradients [C]. 11th International Joint Conference on Computer Science and Software Engineering (JCSSE), 2014Chon BuriIEEE34-39

[11] PengH-y, WangX, WangH-y, YangW-wu. Recognition of low-resolution logos in vehicle images based on statistical random sparse distribution [J]. IEEE Transactions on Intelligent Transportation Systems, 2014, 16(2): 1-11

[12] LiB, TianB, LiY, WenDing. Component-based license plate detection using conditional random field model [J]. IEEE Transactions on Intelligent Transportation Systems, 2013, 14: 1690-1699

[13] YaoZ-j, YiW-dong. License plate detection based on multistage information fusion [J]. Information Fusion, 2014, 18: 78-85

[14] YangA-f, LuM, TengS-h, SunJ-xiang. Local sparse representation for astronomical image denoising [J]. Journal of Central South University, 2013, 20: 2720-2727

[15] OtsuN. A threshold selection method from gray-level histograms [J]. IEEE Trans Syst Man Cybern, 1979, 9: 62-66

[16] WangJ, RenM-wu. Vehicle-logo location method based on prior knowledge and texture directionality [J]. Computer Applications and Software, 2014, 31(5): 163-167