Text Recognition -- Real World Data and Where to Find Them

TL;DR

This paper introduces a method to leverage weakly annotated images for enhancing scene text recognition by generating pseudo ground truth, leading to significant accuracy improvements across multiple datasets.

Contribution

It proposes a novel approach to extract high-quality text instances from weakly annotated data, improving recognition accuracy without requiring fully labeled datasets.

Findings

Achieves a 3.7% average accuracy improvement on benchmarks.

Attains a 24.5% improvement on a weakly annotated dataset.

Produces nearly error-free localized text instances as pseudo ground truth.

Abstract

We present a method for exploiting weakly annotated images to improve text extraction pipelines. The approach uses an arbitrary end-to-end text recognition system to obtain text region proposals and their, possibly erroneous, transcriptions. The proposed method includes matching of imprecise transcription to weak annotations and edit distance guided neighbourhood search. It produces nearly error-free, localised instances of scene text, which we treat as "pseudo ground truth" (PGT). We apply the method to two weakly-annotated datasets. Training with the extracted PGT consistently improves the accuracy of a state of the art recognition model, by 3.7~\% on average, across different benchmark datasets (image domains) and 24.5~\% on one of the weakly annotated datasets.