Abstract
Artificial Intelligence (AI) is the creation of machines that perform functions that require intelligence when performed by humans. AI is currently used in a variety of applications, such as in augmented reality, writing poetry, driving a car on a crowded street, diagnosing diseases etc. In image acquisition devices, an application of AI is the detection and recognition of text and digits in natural scene images. Recognizing natural scene image text is a very challenging area because of the large variability in appearance, with a lot of confusing backgrounds. Natural scene images are being increasingly used in areas like navigation and translation etc., making robust, accurate and fast digit/text recognition systems a necessity. This paper presents a novel DiGiConvNet architecture to read digits in Natural Scene Images. The DiGiConvNet has a unique parallel Convolutional Neural Network (ConvNet or CNN), trained on Maximally Stable Extremal Regions (MSER) image components enhanced using Contrast Limited Adaptive Histogram Equalization (CLAHE). Good results have been obtained on the full natural scene test images of the Street View House Numbers (SVHN) dataset [18].
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
