Block-level Text Spotting with LLMs
This work addresses the need for extracting blocks of text with more context to enhance downstream applications like translation, representing an incremental advancement in text spotting.
The paper tackles the problem of block-level text spotting in images, which is relatively unexplored compared to character, word, or line-level methods, and proposes BTS-LLM, a novel method that uses a large language model (LLM) to group lines into blocks and rectify recognition errors, resulting in semantically meaningful text extraction.
Text spotting has seen tremendous progress in recent years yielding performant techniques which can extract text at the character, word or line level. However, extracting blocks of text from images (block-level text spotting) is relatively unexplored. Blocks contain more context than individual lines, words or characters and so block-level text spotting would enhance downstream applications, such as translation, which benefit from added context. We propose a novel method, BTS-LLM (Block-level Text Spotting with LLMs), to identify text at the block level. BTS-LLM has three parts: 1) detecting and recognizing text at the line level, 2) grouping lines into blocks and 3) finding the best order of lines within a block using a large language model (LLM). We aim to exploit the strong semantic knowledge in LLMs for accurate block-level text spotting. Consequently if the text spotted is semantically meaningful but has been corrupted during text recognition, the LLM is also able to rectify mistakes in the text and produce a reconstruction of it.