Sorry, you need to enable JavaScript to visit this website.

DOES SUPER-RESOLUTION IMPROVE OCR PERFORMANCE IN THE REAL WORLD? A CASE STUDY ON IMAGES OF RECEIPTS

Primary tabs

Citation Author(s):
Vivien Robert, Hugues Talbot
Submitted by:
Vivien Robert
Last updated:
13 November 2020 - 12:52pm
Document Type:
Presentation Slides
Document Year:
2020
Event:
Presenters Name:
Vivien Robert
Paper Code:
2624

Abstract 

Abstract: 

Recently, many deep learning methods have been used to handle single image super-resolution (SISR) tasks and often achieve state-of-the-art performance. From a visual point of view, the results look convincing. Yet, does it mean that those techniques are reliable and robust enough to be implemented in real business cases to enhance the performance of other computer vision tasks? In this article, we investigate the use of SISR to construct higher resolution images of real receipt photos sent by a company’s customers and evaluate its impact on the performance of an OCR task (receipt information retrieval). Using built in task-based performance evaluation methods, we show that the use of SISR can significantly improve OCR performance in the case where recognition was poor in low resolution, but can also deteriorate the performance for receipts that were already successfully recognized. As a conclusion, we provide recommendations on how to best use SISR in a production environment.

up
0 users have voted:

Dataset Files

ICIP Thesis Presentation.pdf

(82)