AON: Towards Arbitrarily-Oriented Text Recognition
This addresses the challenge of scene text recognition for applications in computer vision, offering a novel method for handling irregular text arrangements that existing methods struggle with.
The paper tackles the problem of recognizing irregularly oriented text in natural images by proposing the Arbitrary Orientation Network (AON), which directly captures deep features of such texts and integrates them with an attention-based decoder to generate character sequences, achieving state-of-the-art performance on irregular datasets like CUTE80 and SVT-Perspective while being comparable on regular datasets.
Recognizing text from natural images is a hot research topic in computer vision due to its various applications. Despite the enduring research of several decades on optical character recognition (OCR), recognizing texts from natural images is still a challenging task. This is because scene texts are often in irregular (e.g. curved, arbitrarily-oriented or seriously distorted) arrangements, which have not yet been well addressed in the literature. Existing methods on text recognition mainly work with regular (horizontal and frontal) texts and cannot be trivially generalized to handle irregular texts. In this paper, we develop the arbitrary orientation network (AON) to directly capture the deep features of irregular texts, which are combined into an attention-based decoder to generate character sequence. The whole network can be trained end-to-end by using only images and word-level annotations. Extensive experiments on various benchmarks, including the CUTE80, SVT-Perspective, IIIT5k, SVT and ICDAR datasets, show that the proposed AON-based method achieves the-state-of-the-art performance in irregular datasets, and is comparable to major existing methods in regular datasets.