Skip to main content

Getting Started

SDK IconOCR SDK

HOCR SDK

HOCR SDK provides OCR technology as an SDK that recognizes and detects characters contained in images and PDF files and extracts them as data. Using proprietary deep learning technology, it accurately recognizes printed text, handwritten text, and characters within background images, and enables key-value extraction from unstructured documents.

🔍 Precise Recognition

🧠 Deep Learning Based

📄 Multi-format Support


Key Features​

🧠

Deep Learning Based Recognition

High accuracy with proprietary deep learning technology

📄

Various Document Support

Recognition of printed text, handwritten text, tables, and characters within background images

⚡

High Performance Processing

Optimized processing performance based on CPU/GPU

🔧

Automatic Preprocessing

Automatic correction and smoothing of geometric distortions


Key Features and Benefits​

📋 Key Features

Character Detection and Recognition
Performance optimization in various environments with proprietary deep learning technology

Image Preprocessing
Automatic smoothing of geometric distortions caused by rotation, bending, and camera shooting angles

Handwritten Text and Background Character Recognition
Recognition of printed text, handwritten text, characters within background images, tables, and rows

Unstructured Document Processing
Conversion to structured data through key-value extraction

Multi-format Support
Processing of various document formats including images and PDFs

GPU Acceleration
Fast processing performance with CUDA-based GPU acceleration

🏆 Core Values

  • Accurate Recognition
    High-accuracy character recognition technology based on deep learning
  • Various Environment Support
    Comprehensive support from printed text to handwritten text
  • Real-time Processing
    Fast processing speed through GPU acceleration

💼 Business Benefits

  • Document Automation
    Automation of manual document input tasks
  • Data Extraction
    Extraction of structured data from unstructured documents
  • Cost Savings
    Reduction in document processing costs and time
  • Scalability
    Support for large-scale document processing

Quick Start Guide​

1. Check System Requirements​

Windows System Requirements

  • Operating System: Windows 10 or higher, 64-bit required
  • Compiler: C++17 or higher supported compiler required
  • Processor: i7-8700K or higher recommended
  • Memory: 8GB RAM or higher (16GB RAM recommended)
  • Storage: 4GB or more storage space (SSD recommended)
  • Other: Microsoft Visual C++ 2015-2022 Redistributable Package required
Detailed System Requirements

Check detailed requirements in Supported Operating Systems.

2. License Activation​

License activation follows these procedures:


Detailed License Guide

Check license types and management methods in the License Guide.

3. Install Required Components​

Common Requirements

  • Libtorch Library: Required for OCR SDK usage
  • CUDA 11.6 or higher: Required when using GPU
  • Encoding Settings: Appropriate encoding settings such as UTF-8
Detailed Installation Guide

Check installation methods in Common Prerequisites.

4. SDK Initialization Methods​

Initialization using Integrated Resource File

// Initialize using IMR file
ArgoSDK* sdk = CreateFromImrModelFile("all_module.2024.01.01.cpu.imr");

5. Hello World Example Execution​

Windows Sample Execution

# Execute sample
SampleWindowsMain.exe

# Check required redistributable package installation
# Microsoft Visual C++ 2015-2022 Redistributable Package

6. Basic OCR Processing Example​

#include "ArgoSDK.h"

int main() {
    // Initialize SDK
    ArgoSDK* sdk = CreateFromImrModelFile("all_module.2024.01.01.cpu.imr");
    
    if (sdk != nullptr) {
        // Set image file path
        const char* imagePath = "sample_image.png";
        
        // Execute synchronous OCR
        C_RESULT_CALLBACK callback;
        callback.result_callback = [](const char* result) {
            printf("OCR Result: %s\n", result);
        };
        
        sdk->run(imagePath, &callback);
        
        // Release SDK
        releaseArgoSDK(sdk);
    }
    
    return 0;
}

SDK Components​

Deployment Components​

Assets

  • sample.{yymmdd}.license: License file
  • all_module.{yy}.{mm}.{dd}.cpu.imr: CPU-based integrated resource file
  • all_module.{yy}.{mm}.{dd}.gpu.imr: GPU-based integrated resource file
  • all_module.{yy}.{mm}.{dd}.cpu.TEXT_DETECTION.smr: Character detection CPU module
  • all_module.{yy}.{mm}.{dd}.gpu.TEXT_DETECTION.smr: Character detection GPU module
  • all_module.{yy}.{mm}.{dd}.cpu.TEXT_RECOGNITION.smr: Character recognition CPU module
  • all_module.{yy}.{mm}.{dd}.gpu.TEXT_RECOGNITION.smr: Character recognition GPU module

Include

  • ArgoSDK.h: C API header file

Libs

  • CPU/: CPU-based execution library
  • GPU/: GPU-based execution library
  • opencv/: OpenCV static library

Sample

  • Sample.cpp: OCR SDK usage example
  • MakeUp_cpu.json: CPU JSON initialization settings
  • MakeUp_gpu.json: GPU JSON initialization settings
  • CMakeLists.txt: Build configuration file
Detailed Components

Check all components in OCR SDK Details.