I. Introduction
Correct recognition of vehicle license plates has numerous use cases that include penalizing irresponsible driving and parking, keeping track of vehicles coming in and going out of parking lots, identifying vehicle ownership and so on. While a lot of work on automatic detection and recognition of license plates has been done for decades, there is still a lot of progress to be made on correct Bangla automatic license plate recognition (ALPR). The variance that exists in Bangla license plates owing to existence of different metro and vehicle types, and the scarcity of adequate samples to catch all of that variety makes Bangla ALPR specially challenging. A Bangladeshi company named HeadBlocks is working with the Dhaka Metropolitan Police (DMP) to address this and has developed a commercial Bangla ALPR system with a four stage pipeline that includes license plate detection, character segmentation, character detection, and character recognition. Although the ALPR system developed by HeadBlocks (referred to as HB- ALPR henceforth) has over 96% accuracy as determined by experiments on their confidential test set, HB-ALPR finds it difficult to correctly identify license plates that are more than 30° skewed on the image plane (ip-skewed) or the viewing plane (vp-skewed). Examples of such ip-skewed and vp-skewed license plates are shown in Fig. 1.