Overview

The SaigeVision® Runtime C/C++ SDK is a real-time vision inspection SDK, developed to easily use the SaigeVision® deep learning inference engine in C/C++.

Features of SaigeVision®

• Easy to use even for non-experts in deep learning.
• Excellent vision inspection performance.
• Fast vision inspection speed.
• Supports various vision inspection scenarios.

Funtions of SaigeVision®

Classification Classifies images according to user-defined criteria.	Detection Detects objects or defects in images in the form of boxes.
Segmentation Detects the exact location of defects at the pixel level.	Image Anomaly Detection Learns from normal images to detect areas that differ from the norm.
Advanced OCR Recognizes text within images.	Image Enhancement Enhances low-quality images to high-quality images.

Feature	Description
Classification	Determines whether an image is normal or defective, and if defective, identifies the type of defect.
Detection	Detects the desired defect areas in the image as rectangles. It also identifies the type of defect in the detected area.
Segmentation	Detects the desired defect areas in the image at the pixel level. It also identifies the type of defect in the detected area.
IAD (Image Anomaly Detection)	Determines whether an image is normal or defective and detects defect areas at the pixel level. It can be trained with only normal images, but it cannot identify the type of defect.
OCR (Optical Character Recognition)	Identifies the location of the desired text in the image and recognizes the characters.
IEN (Image Enhancement)	Improves the quality of the image.