Proceedings of the 29th ACM International Conference on Multimedia 2021
DOI: 10.1145/3474085.3478328
|View full text |Cite
|
Sign up to set email alerts
|

Mmocr

Abstract: We present MMOCR-an open-source toolbox which provides a comprehensive pipeline for text detection and recognition, as well as their downstream tasks such as named entity recognition and key information extraction. MMOCR implements 14 state-of-theart algorithms, which is significantly more than all the existing open-source OCR projects we are aware of to date. To facilitate future research and industrial applications of text recognition-related problems, we also provide a large number of trained models and det… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
4
0

Year Published

2022
2022
2024
2024

Publication Types

Select...
5
3
2

Relationship

0
10

Authors

Journals

citations
Cited by 49 publications
(8 citation statements)
references
References 36 publications
0
4
0
Order By: Relevance
“…Text Line Segmentation. We employ PSENet (Wang et al 2019b) and FCENet (Zhu et al 2021) as comparison methods, which are implemented by the mmocr codebase (Kuang et al 2021). By incorporating a progressive scale expansion mechanism and multi-scale kernels, PSENet is able to gradually expand the predicted text-line region, which enables the model to effectively distinguish densely packed text lines.…”
Section: Comparisons With State-of-the-art Methodsmentioning
confidence: 99%
“…Text Line Segmentation. We employ PSENet (Wang et al 2019b) and FCENet (Zhu et al 2021) as comparison methods, which are implemented by the mmocr codebase (Kuang et al 2021). By incorporating a progressive scale expansion mechanism and multi-scale kernels, PSENet is able to gradually expand the predicted text-line region, which enables the model to effectively distinguish densely packed text lines.…”
Section: Comparisons With State-of-the-art Methodsmentioning
confidence: 99%
“…The raw images captured by the camera are of size 1280 × 720. All other hyperparameters and the optimizer, not mentioned below, are set to their default values as specified in MMOCR [47].…”
Section: Datasets and Implementation Detailsmentioning
confidence: 99%
“…In our experiments, the Openmmlab-MMOCR [11] is selected as our toolbox, which is flexible for STD tasks. The hyperparameters of all experiments are set according to the official documentation of Openmmlab-MMOCR.…”
Section: Implementation Detailsmentioning
confidence: 99%