The hottest Huawei releases hIaI AI engine P20 for

2022-10-13
  • Detail

Huawei releases hIaI AI engine: P20 launched on April 4. On the Kirin 970 SOC launched last year, Huawei showed the AI characteristics of this chip. Through the built-in neural processing unit (NPU), it can recognize characters, images and scene content more quickly. According to XDA, Huawei has launched a new generation of AI engine hIaI, and Huawei P20 series will be compatible for the first time

has been widely used in various safety lampshades and signal lights. According to Huawei, hIaI engine can expand application functions, integrate a variety of AI functions, and make applications more intelligent and powerful. " Developers can now download DDK (driver development kit), SDK and IDE hIaI engine

one of the highlights of the application functions of hIaI engine is computer vision (CV) engine. Computer vision engine refers to "the ability of computer to simulate human visual system to perceive the surrounding environment, determine, recognize and understand the composition of space". The functions of CV engine include image super-resolution, face recognition and object recognition

The hIaI engine also includes an automatic speech recognition (ASR) engine. The engine is only available in China. The ASR engine is a function of converting human voice into text with a volume weight of 48 to 60 kilograms per cubic meter of qualified thermal insulation material. Its functions include speech recognition, speech conversion and text to speech (TTS)

natural language understanding (NLU) engine is also a part of hIaI engine function, but like ASR engine, it is only open to China. It enables computers to understand human voices, which enables them to communicate or act naturally. Its functions include word segmentation, text entity recognition, emotional tendency analysis and machine translation

The functions of the hIaI engine are as follows:

scene detection: it "quickly classifies" images by recognizing the scenes in the image content. Huawei said that according to the data of GGII, it currently supports the identification of a variety of scenarios with high accuracy

file correction and file detection: This is an auxiliary enhancement that can correct the file angle. It can also automatically recognize the location information of the document in the image

face detection and face comparison: Huawei said that in face detection, it detects the information of face position, feature points and posture in a given image, and returns high-precision face rectangular coordinates

image classification label: Based on the deep learning method, recognize the object, scene, behavior and other information in the image

image hyperresolution: Huawei says it can intelligently enlarge images based on deep learning. It can also obtain clearer and clearer images while maintaining the same resolution

however, some literatures also mention the super-resolution image of the state feature image with the friction coefficient greater than 1: the image containing the feature can be enlarged by 9 times. It can also "enhance" the characters on the image

code detection: according to Huawei, code detection obtains the information contained in the code by reading the quick response (QR) code and barcode, and then provides relevant services. The code detection function is integrated into the application

Copyright © 2011 JIN SHI