DENK-OCR
Text and character recognition
Fast and precise text and character recognition on all products and objects. Developed for demanding industrial environments.
OCR for every industry
Automated text recognition
OCR, or Optical Character Recognition, is a technology used to convert different types of documents, such as scanned paper documents or images, into editable and searchable text. It recognizes printed or handwritten text in images and transforms it into machine-readable text.
In combination with machines and robots, our technology enables perfected automation processes.
- Text
- Special characters
- Dates
- Numbers
- much more
For our customers, this means optimised use of resources, including time and cost savings, error reduction, barrier-free access and sustainability. Ultra-fast reading for efficient quality control, sorting and administrative work.
- Powerful technology
Also recognizes complex fonts and formats.
- Flexible
In every fallow and intuitive usable.
- Cost-effective
Automation saves time and resources.
Quick start with pre-trained OCR DENKnet
This robustness is reflected in the fact that our technology is stable even in different light situations, when exposed to dirt or even when the marking material is defective. All product labelling can be processed, even in forms that are difficult to read, e.g. glossy die-cuts or blurred embossing.
The solution for automated text recognition
Fast start with pre-trained OCR-DENKnet
With DENK-OCR, you can have printed and legible fonts read out without any AI training. No image preparation is necessary for the use of AI-based image analysis by DENKweit. Our OCR works accurately with legible text without any further training.
For absolute reliability, especially with handwritten, blurred or otherwise difficult-to-read fonts, simply train the network further with a few images. AI training is possible (as with all DENK technologies) in the DENK VISION AI Hub without any programming knowledge. The software can be easily integrated into on-premise or cloud systems.
Which font do you want to read?
We continue where others fail.
Dotted
Shiny metallic
Punched
Arranged in a ring
Blurred font
Embossings
WORKFLOW
From text spotting to OCR
Image capture
The source material is image data from your environment – e.g. from production cameras, scanners or other existing systems.
AI training – Step 1: Identify text areas
During the training phase, the AI learns where in the image it should look for text or characters - text spotting. Technology: object recognition or segmentation.
AI training – Step 2: Recognising symbols and words
In the second training step, OCR is trained: the AI learns what is written in the recognised areas – letters, numbers, symbols, smudged characters, needle-punched, embossed, dotted fonts, etc.
Application of the finished model
In practical operation, DENKnet performs both recognition steps offline and almost simultaneously. → This process is invisible to the user – the corresponding text is read on an image.
Embedded into a sophisticated product
What is behind our reliable OCR?
The ability not only to recognize individual characters, but also to know the relationships between character sequences such as serial numbers or words and to take this knowledge into account during character recognition.
If a customer uses completely unknown characters or print types, DENKnet can be easily retrained. Our DENK VISION AI Hub is designed so that it can be operated without any programming knowledge. The addition of individual images ensures high-performance OCR even in unusual areas of application.
