Basic Concepts

Face Detection

Face detection refers to detecting and marking the area of a person's face from a static image, and giving a rectangular coordinates of a human face (upper left corner coordinates and lower right corner coordinates, wherein the upper left corner of the image is (0,0)), The coordinates of the lower right corner are (n, m), and n and m are respectively the pixel width and height of the image). The face and gender information is added to the 0.2, 0.3 version of the service.
Face detection has a wide range of usage, such as (1): Intelligent driver assistance system, which can automatically recognize the face in the front of the road and prompt the driver to pay attention to pedestrian in front (2): Adding face detection algorithm to camera can real-time capture the position of the face in the frame and auto-focus.

Face Recognition

The premise of face recognition is the ability to detect the presence of a human face. And then identify whether the detected face is of someone you know. This involves the use of human face contrasts, comparing similarities of two faces from various angles.
The most typical application for face recognition is face recognition payment and face recognition access. The user only needs to provide some negatives to the system (the user's own photo showing a clear face), and can pay, access control and so on through the next scan on the face.

Object Recognition

Given a static image, the system can analyze the objects presented in the photo. As accuracy rate cannot reach 100%, the confidence of the corresponding judgment is also provided.

Natural Language Machine Translation

A machine translation system is an online service that supports translating a large amount of text; it is a process of translating text from the "source" language to a different "target" language. Since the early 2010s, the new AI technology, deep neural networks (also known as Cloud-ML(应该是deep-learning)) have helped achieve high translation accuracy. It now supports 7 languages for mutual translations.

OCR Text Recognition and Image Translation

OCR (optical character recognition) refers to a service using electronic device (such as scanner or digital camera) to check for printed characters on paper, and then using a character recognition method to convert the shapes into searchable, processable, and translatable text. We also provide services that directly translate from an image containing the source language to be translated into a target language, with positional coordinate information provided.