# Full-Page Text Recognition: Learning Where to Start and When to Stop

## TL;DR

This paper introduces a novel full-page text recognition method that uses CNNs and LSTMs to efficiently localize text lines by predicting only their starting points, improving recognition in diverse documents.

## Contribution

It proposes a new approach combining CNNs and LSTMs for localized text line detection by predicting only line start positions, enhancing efficiency and accuracy.

## Key findings

- Effective on heterogeneous Maurdor dataset
- Improved localization accuracy
- Reduced computational complexity

## Abstract

Text line detection and localization is a crucial step for full page document analysis, but still suffers from heterogeneity of real life documents. In this paper, we present a new approach for full page text recognition. Localization of the text lines is based on regressions with Fully Convolutional Neural Networks and Multidimensional Long Short-Term Memory as contextual layers. In order to increase the efficiency of this localization method, only the position of the left side of the text lines are predicted. The text recognizer is then in charge of predicting the end of the text to recognize. This method has shown good results for full page text recognition on the highly heterogeneous Maurdor dataset.