摘要
Tracking images using shape descriptor can be more accurate than using other existing methods and it is most useful when the environment is complex. However the existing methods with shape descriptor get more labeled parts to compare and detect the object in an image, which makes the computation more complicated. Thus, we need a trade-off between the accuracy and efficiency requirements. This paper aims to bridge this gap between the accuracy and efficiency requirements by using morphology method. To improve the original monochromatic object detecting system, we propose a new color descriptor to preprocess the image with polychromatic object. Experiments have been conducted and shown the proposed method has made a great improvement in the time complexity minimization comparing with the performances of the original detection algorithm.
Tracking images using shape descriptor can be more accurate than using other existing methods and it is most useful when the environment is complex. However the existing methods with shape descriptor get more labeled parts to compare and detect the object in an image, which makes the computation more complicated. Thus, we need a trade-off between the accuracy and efficiency requirements. This paper aims to bridge this gap between the accuracy and efficiency requirements by using morphology method. To improve the original monochromatic object detecting system, we propose a new color descriptor to preprocess the image with polychromatic object. Experiments have been conducted and shown the proposed method has made a great improvement in the time complexity minimization comparing with the performances of the original detection algorithm.
