Effect of foreground and background occlusion onfeature matching for target location
E. Roy Davies · Electronics Letters · 1999
The problem of designing optimal templates for detecting image features is discussed. In particular, the factors determining optimal mask size are examined. It is found that noise and occlusion are important relevant factors, and formulas are presented which summarise the situation. Curiously, variable occlusion of the background could be one of the most important problems in target location and related tasks.