Video Object Matching Based on SIFT and Rotation Invariant LBP

Deng Yi, Lu Jianguo, Qu Xilong


Object detection and tracking is an essential preliminary task in event analysis systems (e.g. Visual surveillance).Typically objects are extracted and tagged, forming representative tracks of their activity. Tagging is usually performed by probabilistic data association. However, as data may have been collected at different times or in different locations, it is often impossible to establish such associations in systems capturing disjoint areas. In this case, appearance matching is a valuable aid. This paper proposes a object matching method for multi-camera by combining HOG and block LBP, and computes accuracy rate by SVM. Using independent tracks of 30 different persons, we show that the proposed representation effectively discriminates visual object and that it presents high resilience to incorrect object segmentation and illumination. Experimental results show that the average accuracy




object matching; multi-camera; non-overlapping views; HOG; LBP

Full Text:



  • There are currently no refbacks.

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

shopify stats IJEECS visitor statistics