Multi-class detection
Detect cars, trucks, buses, motorcycles, and multiple vehicles in one image.
GTOcr Recognition Engine
AI recognition enginefor vehicle and plate OCR
GTOcr combines vehicle detection, plate localization, and OCR into one ANPR pipeline for parking management, access control, and traffic enforcement.
- 99%+
- Plate accuracy
- < 100ms
- CPU processing
- KR/TH
- Korean and Thai plates
Recognition pipeline for industrial sites
The engine returns structured vehicle boxes, plate regions, OCR text, and confidence scores.
Plate localization
Separates plate candidates from vehicle regions and calculates recognition-ready areas.
Deep learning OCR
Select Korean or Thai plate models for country-specific recognition.
JSON output
Returns text, boxes, and confidence scores in a format ready for integration.
Connect results directly to control software
Recognition results can feed barrier control, visitor checks, permit validation, and entry logs.
- 1Image input
- 2Vehicle detection
- 3Plate OCR
- 4System integration
Fast integration with four functions
Delivered as a Windows x64 Native DLL, GTOcr can be integrated into C, C++, C#, Python, and Delphi environments.
CPU-first
Works on servers and edge devices without GPU dependency
Multi-country
Korean and Thai plate model support
Native DLL
Low-friction integration with existing software
Field-ready
Result data shaped for parking, access, and enforcement
GTOcr API flowWindows x64 DLL
01
Init
02
Inference
03
Result
04
Final
{
"class": "vehicle",
"conf": 0.95,
"plate_info": [{
"text": "86구0623",
"conf": { "ocr": 0.99 }
}]
}Check recognition performance in the live demo
Upload a vehicle image and review recognized plates together with the raw JSON result.