Fast recognition

Very simple and clear images, such as high-quality scans, do not require automated corrections and areas detection.

Aspose.OCR for C++ can work in the fastest recognition mode that consumes minimum possible resources using asposeocr_fast_recognize function. This function takes one or more OcrInput objects and returns a list of recognition results, one result per image.

// Provide the image
string file = current_dir + "/source.png";
AsposeOCRInput source;
source.url = file.c_str();
std::vector<AsposeOCRInput> content = { source };
// Extract text from the image
auto result = asposeocr_fast_recognize(, content.size());
// Output the recognized text
wchar_t* buffer = asposeocr_serialize_result(result, buffer_size, export_format::text);
std::wcout << std::wstring(buffer) << std::endl;

Performance impact

This function is about twice as fast as regular recognition and is recommended for public web applications and mobile devices.


This recognition mode does not work with skewed images and does not support recognition settings. However, you can preprocess an image before sending it to the OCR engine.