Abstract | ||
---|---|---|
Modern CRNN OCR models require a fixed line height for all images, and it is known that, up to a point, increasing this input resolution improves recognition performance. However, doing so by simply increasing the line height of input images without changing the CRNN architecture has a large cost in memory and computation (they both scale O(n
<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>
) w.r.t. the input line height).We introduce a few very small convolutional and max pooling layers to a CRNN model to rapidly downsample high resolution images to a more manageable resolution before passing off to the "base" CRNN model. Doing this greatly improves recognition performance with a very modest increase in computation and memory requirements. We show a 33% relative improvement in WER, from 8.8% to 5.9% when increasing the input resolution from 30px line height to 240px line height on Open-HART/MADCAT Arabic handwriting data.This is a new state of the art result on Arabic handwriting, and the large improvement from an already strong baseline shows the impact of this technique. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1109/ASAR.2018.8480182 | 2018 IEEE 2nd International Workshop on Arabic and Derived Script Analysis and Recognition (ASAR) |
Keywords | Field | DocType |
fixed line height,input resolution,recognition performance,input images,CRNN architecture,computation,input line height,convolutional pooling layers,max pooling layers,manageable resolution,base CRNN model,memory requirements,neural OCR models,CRNN OCR models,downsample high resolution images | Decimation,Pattern recognition,Arabic handwriting,Computer science,Pooling,Artificial intelligence,Computation | Conference |
ISBN | Citations | PageRank |
978-1-5386-1460-0 | 1 | 0.41 |
References | Authors | |
0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
stephen rawls | 1 | 59 | 4.08 |
Huaigu Cao | 2 | 347 | 29.09 |
Joe Mathai | 3 | 3 | 0.79 |
Premkumar Natarajan | 4 | 874 | 79.46 |