An adaptive tracking algorithm based on particle normalization operation
| Autor: | Ge Guangying, Tian Cunwei |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Computer science
business.industry Normalization (image processing) 020206 networking & telecommunications 02 engineering and technology 01 natural sciences 010309 optics Atmospheric measurements Kernel (image processing) 0103 physical sciences 0202 electrical engineering electronic engineering information engineering Bhattacharyya distance Computer vision Artificial intelligence Adaptive tracking Particle filter business Algorithm Eigenvalues and eigenvectors |
| Zdroj: | 2017 2nd International Conference on Image, Vision and Computing (ICIVC). |
| DOI: | 10.1109/icivc.2017.7984593 |
| Popis: | Based on bug of traditional Particle Filter tracking algorithm, a new adaptive tracking algorithm is presented using particle normalization operation. The oval tracking template is used in this algorithm, and the center coordinates, semi-major axis and angle are selected as eigenvector of the particle. The image normalization operators are used to process the sample particle. The relevance of the target template and area to be test are calculated by the Bhattacharyya coefficient. And the weights of particles are calculated in the recurrence. Experiment results show that this algorithm can track object accurately and efficiently when the object change in size and angle. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|