Thresholding based on learning theory

Learning theory is applied to real-time thresholding of gray-level images on a line-by-line basis. With this approach, each scan line becomes an isolated entity and a dynamic threshold value for each pixel is iteratively calculated while scanning. The approach works with horizontal or vertical scanning. The locations of objects in the image may also be identified while thresholding. This approach can be used to develop fast real-time algorithms suitable for hardware implementation