… text in natural scenes, known as scene text recognition (STR), is regarded as a core task of
optical character recognition … Recently, we observe that visual-language models like CLIP [48]…

… [13] propose to ensemble attention and language models in an attention-based architecture.
… Images are generated from side-view angle snapshots in Google Street View. Therefore, …

… (b) Scene text recognition mistakes in the STVQA task. … out the correct scene text words, we
design a language refinement network based on a pre-trained language model to distinguish …

… scene text recognition approach by tuning a vision and language reconstructor to a text …
Our approach, based on mask and reconstruction, not only learns rich visual and semantic …

… recognition accuracy remains far from expectations. The Transformer’s architecture [25] is a
novel framework, introduced first for natural language … exist in natural language processing …

… to most deep learning based text recognition models. 2. We design a novel scene text
recognition algorithm that learns a descriptive and robust text representation (image feature) …

… Scene text recognition is a challenging task due to diverse variations of text instances in
natural scene … efficient feature representations for multi-oriented scene texts. In this paper, we …

… a new model for scene text recognition that only consists of … scene text recognition network
that does not use any building blocks especially designed for the task of scene text recognition…

… visual features and language … language Tamil, Malayalam, and Telugu scene text images
that are part of the IIIT-ILST dataset. The scripts for these languages share similar character …

… Finally, for the VM in scene text recognition, we develop a position attention module to make
visual predictions parallelly based on character order, which employs a U-Net to enhance …