MMU-OCR-21: Towards End-to-End Urdu Text Recognition Using Deep Learning

Abstract

Optical Character Recognition (OCR) is a technique that generates text from an image. Recognizing the importance of OCR in real-world settings, a plethora of techniques have been developed for Western, as well as Asian languages. Urdu is a prominent South Asian language and a number of different solutions for Urdu OCR have been proposed. However, fewer attempts have been made to develop end-to-end deep learning-based solutions for recognizing printed Urdu text. Furthermore, several benchmark corpora for Urdu OCR have been developed that can be used for training and evaluation of different OCR techniques. However, there are a number of limitations of the existing Urdu corpora: firstly, most of them have either character or word or text images, which are usually rendered using only a single font, Nastaleeq. Secondly, the volume of the existing datasets is so small that it is not suitable for working with the deep-learning techniques that have achieved groundbreaking results for OCRs. To that end, in this study, we have proposed a very large Multi-level and Multi-script Urdu corpus (MMU-OCR-21). It is the largest-ever Urdu corpus of printed text that is effectively suitable to work with deep learning techniques. In total, the corpus is composed of over 602,472 images, including text-line and word images in three prominent fonts, and their respective ground truth. Also, we have performed experiments using multiple state-of-the-art deep learning techniques for text-line and word level images.

Hybrid manifold smoothing and label propagation technique for Kannada handwritten character recognition

“…The episodic framework was introduced by Nasir et al (2021). It provides a simulation for training a meta-learning model for few-shot classification tasks.…”

“…The challenge of converting manuscripts and printed documents into digital formats has been the focus of computer vision research (Nasir et al, 2021;Gowda and Kanchana, 2022). Recent advances have blurred the interface between physical copies of text and their digital counterparts.…”

Enhancing Optical Character Recognition on Images with Mixed Text Using Semantic Segmentation

“…The architecture is essentially an end-to-end fully convolutional network (FCN), and in essence, it only contains convolutional layers and no dense layers, enabling it to be used with images of varying sizes [24]. In Figure 8 given input of an image of a hand-filled form, a trained semantic segmentation algorithm can output a segmentation map, a blank image wherein only the areas of interest (in this case, relevant text) are highlighted in black [25]. This segmentation map is crucial in understanding which areas of an image pertain to the major class in consideration.…”

Komparasi Metode Backpropagation Neural Network dan Convolutional Neural Network Pada Pengenalan Pola Tulisan Tangan

“…Pengimplementasian OCR dapat dilakukan diberbagai ranah penelitian seperti, structural text form recognition [6], automatic number plate recognition [7], alat bantu baca tuna netra, pengenalan tulisan tangan (handwritten) dan sebagainya. OCR juga dapat digunakan untuk melestarikan dan mendigitalkan script [3] atau sastra kuno sebagai warisan [1] [8]. Identifikasi script merupakan langkah penting dalam pemrosesan gambar dokumen terutama ketika lingkungan multiskrip atau multibahasa [9].…”