TEACHING HANDWRITTEN DOCUMENT RESTORATION TECHNIQUES USING STATISTICAL PROCESSING

Abstract

We digitalized a corpus of handwrittendocuments from an Heritage Preservation Program. Severaldocuments were severely damaged due to time and storageenvironment. In a traditional image processing course,several filters will be applied to recover the images fromselected degradations. To understand these filters, studentsmust have had an exposure to signal analysis techniques. Wedeparted from the traditional scheme using statistical imageanalysis techniques. Students were able to perform somemeasurement of the degradations and their recovery withoutrestoring to signal analysis concepts. The degraded image isstatistically modeled and the observation probability densityis obtained from the gray-level histogram of the input image.The use of historical documents motivated our students asthey were able to recognize in the document historicalsituations they have studied in the high school.Index Terms ⎯ Image Processing, Introduction Courses,Classroom Assessment Techniques, Math Awareness