Learning to Read by Spelling: Towards Unsupervised Text Recognition
It addresses the need for unsupervised text recognition to avoid large aligned datasets, which is incremental as it builds on existing alignment methods for a specific domain.
This work tackles the problem of visual text recognition without paired supervisory data by aligning predicted strings from text images with lexically valid strings from corpora, achieving excellent accuracy on synthetic and real printed book images without labeled examples.
This work presents a method for visual text recognition without using any paired supervisory data. We formulate the text recognition task as one of aligning the conditional distribution of strings predicted from given text images, with lexically valid strings sampled from target corpora. This enables fully automated, and unsupervised learning from just line-level text-images, and unpaired text-string samples, obviating the need for large aligned datasets. We present detailed analysis for various aspects of the proposed method, namely - (1) impact of the length of training sequences on convergence, (2) relation between character frequencies and the order in which they are learnt, (3) generalisation ability of our recognition network to inputs of arbitrary lengths, and (4) impact of varying the text corpus on recognition accuracy. Finally, we demonstrate excellent text recognition accuracy on both synthetically generated text images, and scanned images of real printed books, using no labelled training examples.