copy-writing - open to bidding

In any OCR system, preprocessing is the primary step that plays an important role in reliable segmentation, feature extraction, and finally recognition. Skew detection/ correction as a preprocessing technique, is widely studied/employed for enhancement of different OCR results [1]. However, in an unconstrained handwritten text-page written in any language especially Arabic and all similar scripts every text-line may not be only skewed, it also would suffer a slow varying oscillation (non-linear alignment). Moreover, a skew detection/correction technique, which results in a global orientation for entire text-page/text-line, cannot be utilized in case of an oscillatory text-line as a preprocessing operation.

.NET Σχεδιασμός Γραφικών HTML PHP Σχεδιασμός Ιστοσελίδας

Ταυτότητα Εργασίας: #5185698

Σχετικά με την εργασία