Natural scene character detection and recognition method based on CRAFT and SCRN-SEED frameworks
A natural scene and text detection technology, applied in the field of natural scene text detection and recognition, can solve the problems of inaccurate text positioning, difficult to accurately locate and mark, and unsuitable text, and achieve the effect of improved accuracy and good effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment Construction
[0031] The natural scene text detection and recognition method based on CRAFT and SCRN-SEED framework described in the present embodiment comprises the following steps:
[0032] (1) Create an image dataset using real datasets and synthetic datasets, divide the image dataset into a training set and a test set, adjust the size of each picture in the image dataset, and convert the image format in the dataset to mdb format.
[0033] The real data sets come from ICDAR2013, ICDAR2015, ICDAR2017, MSRA-TD500, TotalText, CTW-1500, and the synthetic data set is SynthText data set.
[0034] (2) Use the image data set to train the CRAFT network, the flow chart is as follows figure 1 shown.
[0035] (201) Improve the CRAFT network, use the ResNet50 network as the backbone network, input the pictures in the synthetic data set to the improved CRAFT network for feature extraction, and output the region score Region Score and affinity score Affinity Score;
[0036] (202) Encoding by Gaussian...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com