Constrained Monotone Regression and Outlier Detection for Searching Occlusion Objects

In this paper, we propose an outlier detection algorithm for searching occluding objects due to moving objects using two images, which are captured with the same scene at different time. In order to reduce the influence from the different intensity properties of the image pair, an intensity compensation scheme, which is based on the polynomial regression model, is employed. The constrained monotone regression is performed to obtain a monotonically increasing compensation function and a scaled residual is considered to detect possible outliers. A backward search algorithm is developed to estimate the residuals. Numerical results for real images show a robust detection performance for various intensity conditions.