Reading Skewed Images without Image Rotation

Aims: To extract characters from skewed images without image rotation.
Study Design: This study is designed to be implemented at the gates of Customs, Port Authorities, Terminal Operators and it can also be implemented for vehicles traffic management.
Place and Duration of Study: Lebanon, between September and November 2013.
Methodology: The proposed method consists of sorting the segmented characters according to the X axis, then assigning a Program Line Number to each character based on the skew angle and finally sorting the Program Line Numbers according to their intersection with the Y axis.
Results: Our approach is capable of handling any font and size of characters and it is robust and efficient; regarding its complexity for an image having N lines and M characters, the worst CPU time usage and the worst memory usage is equal to O (NxM) while the network usage and disk usage for one image is O (1) which led to a 0.11 milliseconds response time to extract all container number digits.
Conclusion: Acceleration of segments’ extraction from skewed images by avoiding image rotation in order to acquire a faster and more accurate OCR process.