The Human License Plate Recognition (HLPR), designed to reduce the costs of manual review through skill-based routing, is a module of riteVision™, our comprehensive machine-learning based vehicle identification and automatic license plate recognition platform that also includes Automatic License Plate Recognition (ALPR) system. HLPR uses characteristics from the machine learning system to route images to the correct reviewer (based on their experience and recorded accuracy), thereby minimizing review times. To optimize to the minimal level of review, the system will route to a particular type of review based on image confidence. There are four types of manual review:
Our riteVision™ HLPR application is designed based on targeted image review-specific UI/UX research, so the application is intuitive and ergonomic, aiming to speed up reviews, increase accuracy, and minimize eye strain.
Our manual image review is performed and supervised by our experienced operations team in Richardson, Texas. ETC’s image review systems process over a million images daily and our operations team manually reviews over 350k images per day, as well as perform the daily audit and quality assurance (QA) activities necessary to maintain performance requirements. For our manual image review services, we currently use internal and external resources both U.S.-based and international on a 24/7 basis. This allows for our daily operations to remain consistent and balance out the negative effects of processing surges and outages.
ETC uses a unique approach to workforce management for image review operations. Using a metric driven process, we compile key statistics from operations to facilitate both workforce management and performance evaluation. Through continuous monitoring of existing operations, we effectively forecast and trend volumes, and personnel requirements, and accommodate accordingly. Audit analysis identifies outliers such as under-performers or staff with excessive corrections. This data, combined with our training and evaluation programs, is used to implement process improvement to satisfy performance and production goals.
In our current image review operations, we perform a daily QA of the reviewers to ensure that performance standards are met. Based on these scheduled audits, we typically provide image review services with less than 1% error rate.