Fast recognition

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

Aspose.OCR can work in the fastest recognition mode that consumes minimum possible resources using the following methods:

Method	Description
page_fast()	Quickly read the image without automatic text areas detection and skew angle correction.
page_fast_from_raw_bytes()	Quickly read the image provided as an array of pixels without automatic text areas detection and skew angle correction.

std::string image_path = "../Data/Source/Source.img";
const size_t len = 4096;
wchar_t buffer[len] = { 0 };
aspose::ocr::page_fast(image_path.c_str(), buffer, len);
std::wcout << buffer;

Performance impact

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

Drawbacks

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.