Abstract
Pattern recognition refers to the process of identifying patterns in data. It deals with automatic discovery of regularities in data using various machine learning algorithms and take actions based on the patterns, like classifying data into appropriate categories, extracting information from image, etc. The data used for pattern recognition can be of different forms: text, images, sound, etc. Pattern recognition has a vast area of applications in the domain of NLP and computer vision. It is used along with data analytics to build forecasting systems. It is used to perform text analysis, plagiarism detection, translation systems functioning, and text correction systems functioning. In face detection, visual search engines use pattern recognition to perform image pattern recognition. Optical character recognition (OCR) refers to analysis of image and converting the images containing alphanumeric information into corresponding machine-encoded text. It uses an algorithm that applies a library of patterns and compares them with available input training data to map the corresponding text in an image, which is then further assessed using relevant language corpus to perform text recognition with reasonable accuracy. An important use case of the above-described OCR for pattern recognition is in automatic number plate recognition (ANPR) systems. This chapter discusses different methods used for pattern recognition with focus on deep learning. The ANPR technology automatically recognizes license plate numbers using OCR and other computer vision methods. Steps involved in implementing an ANPR system for security and surveillance are discussed in detail. Both traditional image-processing techniques and deep learning model were implemented to compare the efficiency and analyze the challenges in improving system accuracy.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have