Multilevel hit-miss transform for object detection

The rank-order hit-miss transform (HMT) filter is a significant new advancement in pattern recognition. We detail a new version of our HMT algorithm used for detection, the filter parameters used, and detection (PD) and false alarm (PFA) results. In detection, these filters are required to locate all objects in a scene with clutter present. This must be achieved for objects in multiple different classes, with 3-D distortion and contrast differences present. Thus, they represent considerably new image processing filters.