Deep learning-based Optical Character Recognition

Deep OCR – reading like a human

Deep OCR is a deep learning-based approach to Optical Character Recognition (OCR) that brings machine vision one step closer to human reading ability.
Compared to existing OCR algorithms, Deep OCR can localize characters much more robustly, regardless of orientation, font, or polarity. By automatically grouping characters, entire words can be identified. This significantly improves recognition performance, as it avoids misinterpretations of similar characters.

DEEP OCR

Advantages

Robust Recognition: Deep OCR localizes characters more precisely, even with different fonts and rotations.
Word Identification: The automatic grouping of characters allows the recognition of entire words, greatly improving performance compared to traditional OCR methods.
Improved Stability: Deep OCR can handle large images and offers better overall stability.
Additional Character Set Support: Deep OCR works with a wide range of characters and fonts, enabling a broader range of applications.
Confidence Score: For each recognized character, a confidence value is calculated, improving recognition accuracy and allowing further refinement of the results.

Deep OCR Training

With HALCON 22.05, MVTec introduced Deep OCR: a training feature that allows users to adapt the technology to their specific applications. The training enables the creation of custom datasets for OCR recognition, addressing rare or specialized fonts as well as challenging text conditions, such as low contrast.

Deep OCR Without Training

Ready to use without training effort

Excellent recognition of standard texts with common fonts and layouts.
Quick start for typical OCR applications.

Tire with number and text recognition –correctly recognized characters with Deep OCR.

Deep OCR with Training

Adaptation to Specific Applications

High-precision recognition of low-contrast texts, such as on tires or challenging surfaces.
Advanced training capabilities for rare or specialized characters and print styles.
Refined model performance for specific OCR tasks.

Advantages Of Deep OCR Training

Training on User-Specific Data: Create custom training datasets for text reading applications.
Difficult Texts: Ideal for hard-to-read texts with low contrast (e.g., tire labels).
Special Fonts: Rarely used special characters and print styles can also be easily trained.

HOW DOES IT WORK

Helpful Tutorials

Please note: Once you watch the video, data will be transmitted to Youtube/Google. For more information, see Google Privacy.

Availability in HALCON & MERLIC

Deep OCR is available in both HALCON and MVTec MERLIC.

Deep OCR training is currently only included in HALCON. Simply label your data in the MVTec Deep Learning Tool and seamlessly integrate it into HALCON.
Learn more about HALCON

MERLIC enables user-friendly use of Deep OCR without requiring deep programming knowledge.
Learn more about MERLIC

OUR KNOWLEDGE & SERVICES

Benefit From Our Expertise

PRACTICAL INSIGHTS & EXPERT KNOWLEDGE

Discover Our White Papers!

Explore our white papers to gain practical insights and expert knowledge on industrial machine vision. Download them now to understand key technologies, current trends, and real-world applications that support informed technical and strategic decisions.

To the white papers

INDIVIDUALLY TAILORED

Evaluation Of Your Application

Do you want to know if we offer the right solution for your industry? Send us your software application design and our experts will check it.

Learn more