An Object Detection Method Based on the Separability Measure using an Optimization Approach

The detection of an object in an image can be set up as a maximization problem using separability measure as the objective function. The parameters of this objective function are the parameters used to define two regions in a mask. This mask has the same dimension as the image being considered. The two regions in the mask correspond to two regions in the image under investigation. The pixel value information in these two regions in the image would be used for the calculation of the separability measure. An optimization method would then be used to solve this maximization problem. This study demonstrated that the proposed method could detect the rectangle in the test image successfully. It showed that object detection and detailed segmentation could be carried out at the same time if the geometry of an object could be described mathematically.