Printed Character Recognition (PCR)

Introduction  Isolated Character Classifier  Constrained Printing Recognition  Unconstrained Printing Recognition  Back to OCR Technologies Homepage   Introduction The recognition system consists of two main processing units — a character separator and an isolated character classifier. Character separation (frequently called segmentation) can work in two modes: fixed (constrained) spacing mode (where character size is known in advance and therefore segmentation can be very robust) variable (arbitrary) spacing (where no a priori information can be assumed) Hence, our demo consists of three parts: recognition of isolated characters, work with constrained segmentation, and work with unconstrained segmentation.   Isolated Character Classifier The recognition module gets on input an extracted and size-normalized image representing a character to be recognized. The module produces on output an ordered list of a few of the most probable classification candidates, together with their confidence values. The task is perfomed by matching the raster sample with template masks representing different characters. The masks are prepared by an off-line training phase. A mask can be considered as a raster image containing three types of pixels: black, white, and undefined (gray). Initially, template masks are built per font. In a single font-set of masks, every character is represented by exactly one mask. Images representing template masks built for Courier font are presented below. In practice, a font character to be recognized is often unknown a priori. Hence, templates representing the most prevalent fonts are prepared and combined together. The Omnifont recognizer, containing a number of masks per character, is shown below. An input image is correlated with all the masks stored in the recognizer. The mask which has the highest correlation score is taken to be the primary result of the recognition. --->--->B(0.99)   Constrained Printing Recognition In this case, character spacing is fixed. Hence, segmentation is possible even when fields are distorted, as illustrated below. Field extracted from medical insurance forms (USA), recognized as Omnifont. Credit card numbers impressed through a copy paper on transaction vouchers.   Unconstrained Printing Recognition In the following example, we can see the main steps of the recognition process . The input image used in the example was extracted from a fax cover-sheet. Possible slant is estimated and compensated (in order to cope with italics and backslanted fonts). Top and bottom base lines are detected. (The base lines are shown in red and blue colors on the following picture.) The whole image is divided into horizontally separated "words." (Two such words are detected in our example. They are separated by a vertical green line.) Each word is processed separately, and decomposed into connected components. The following is the decomposition of the first word:                                      The connected components undergo further analysis. Some of them are decomposed into smaller parts (we call them atoms).                                                                                         Thus, the problem of characters separation is reduced to a problem of correct partition of an ordered sequence of atoms. In other words, we need to combine the atoms to molecules. Of course, this can be done in a variety of ways. Three possible molecule sequences are shown below: Segmentation #1 -- -- -- -- -- Segmentation #2 -- -- -- -- -- -- Segmentation #3 -- -- -- -- -- -- -- But how do we choose among all the possible partitions the "correct" one shown below? Correct Segmentation -- -- -- -- -- -- This choice is performed by using the recognition confidence values, produced by the character classification kernel described above. All the molecules are recognized separately. The average value of the recognition probabilities obtained for the corresponding molecules provides an estimate of the confidence of the entire word. For the above examples: Segmentation #1 - confidence = 0.83 Segmentation #2 - confidence = 0.79 Segmentation #3 - confidence = 0.71 Correct segmentation - confidence = 0.95 Final segmentation and recognition result: This process enables the successful recognition of broken and connected characters and dot-matrix printing.