Trier and Jain: Goal-directed Evaluation of Binarization Methods

| This paper presents a methodology for evaluation of low-level image analysis methods, using binariza-tion (two-level thresholding) as an example. Binarization of scanned gray scale images is the rst step in most document image analysis systems. Selection of an appropriate binarization method for an input image domain is a diicult problem. Typically, a human expert evaluates the binarized images according to his/her visual criteria. However, to conduct an objective evaluation, one needs to investigate how well the subsequent image analysis steps will perform on the binarized image. We call this approach goal-directed evaluation, and it can be used to evaluate other low-level image processing methods as well. Our evaluation of bina-rization methods is in the context of digit recognition, so we deene the performance of the character recognition module as the objective measure. Eleven diierent locally adaptive binarization methods were evaluated, and Niblack's method gave the best performance.