ZTRUS we are the owner of AI-OCR
Our technology can transform data from images of documents with different layouts into expected digital formats without fixed templates or fixed keywords but extracting information via cognitive models learned from pre-trained data. ZTRUS’s technology is also capable of continuously learning and updating its model throughout its usage lifespan to maintain the high-level of quality and efficiency of its operations.
We possess an AI technology that serves as a Matching Module, which is responsible for pairing data obtained from the conversion of images into text with datasets that share the same meaning. This technology continually learns and evolves to enhance accuracy and precision, achieving higher levels of correctness.
Key Highlight
- Extract data from image files into text using OCR technology, supporting both Thai and English languages for now and having other languages in the product pipeline.
- Pre-trained Extraction : Capable of extracting data from documents with various layouts and font styles, including standard computer fonts or those from typewriters and printers with pre-trained cognitive models rather than fixed templates or keywords.
- Data Matching with NLU (Natural Language Understanding): High accuracy in Data Matching by the meaning or substance rather than spelling.
- AI : The capabilities of artificial intelligence technology include notifying, verifying, screening, and assessing data, along with customizable algorithms and parameters.
- Machine Learning : Our technology is designed to continuously learn and improve to maximize the performance for each customer. #6AA768
Support more than 26 document types
- Invoice and Receipt
- PO
- Quotation
- Utility bill
- Electricity & Water bill
- Phone & Internet bill
- Express Bill
- Hospital Invoice
- Shipping invoice
- BL / Airway Bill
- Custom Document
- Withholding Tax Document
- Bank statement
- Bank book
- Pay slip
- NCB Document
- ID Card & Passport
- House registration
- Business registration
- Insurance claim
- Application form
- Car insurance policy
- Driver license
- Government Document
- Contracts & legal documents
- Other documents ex. Business-specific documents (Able to customize AI models to process business-specific documents as well)
Technology AI-OCR
Cognitive-based Candidate Selection
Step 1
We "Read" the document using our in-house developed OCR module which can work "on local" and "on public"
Step 2
We "Analyze" their structures, patterns, and potential meanings, then We
Step 3
We "Generate" candidates and "Select" the most suitable ones based on their confidence scores from our Cognitive Extraction.
Not fixed template/Keywords
Generation of OCR Technology
This is where we have been Deep Learning OCR
Traditional OCR
Concept of Technology : Guessing Alphabet from Image
Deep Learning OCR
Concept of Technology : Guessing Words from Image with surrounding alphabets
LLM-OCR
Concept of Technology : Guessing Words from Image with surrounding Words
Our strengths
support various document types
The AI-OCR system can process a wide range of document formats, both fixed formats like identification cards, driver's licenses, and varied formats like invoices and receipts, as well as other types of documents.
Automated data entry into the system
Enhance the efficiency of document management across various departments within the organization, resulting in significantly faster workflow processes and time savings for employees. This is achieved by leveraging OCR technology empowered by AI for document processing.
support documents in both Thai and English languages
The AI-OCR can read documents in both Thai and English languages, including the capability to read Thai numerals.
Data Extraction from tabular image
The AI-OCR system can instantly extract data from various documents presented in table format and present it as a table such as excel, csv and json without the need to pre-create templates. This caters to businesses seeking convenience, speed, and automated processes in their operations.
Continuously learn and evolve
Can learn to enhance operational efficiency, including the ability to read images with low sharpness or insufficient brightness.
High accuracy results in a reduction of workload by over 80%
We utilize AI and machine learning algorithms for data transformation, achieving high accuracy and reducing workload by over 80%.
Happy Clients
Projects Done
Days Of Work
Award Winner
Consult with Our Experts
Our team is available to offer comprehensive consultation on AI-powered OCR solutions, including functionality, integration options, pricing structures, and other related aspects.