Recently, AI technology’s application covers many fields, from smart security to autonomous driving. And behind every achievement is inseparable from strong data support. As the core factor of AI algorithm, datasets aren’t just the basis for model training, but also the key factor for improving mode performance, By continuously collecting and labeling various datasets, developer can accomplish application with more smarter, efficient system.

On June 29 2021, Huawei published a new patent for smartphone unlocking, entitled “equipment unlocking method and device”. The patent abstract shows that this application can use the biological attributes of lip movement characteristics for identity authentication, thereby unlocking the device, and has flexibility, anti-forgetting, anti-plagiarism, and modifiability.

The unlocking method consists of a scanning unit and an identifying unit. Just like face unlock we are familiar with, the device uses pre-store lip movements to perform the authentication. Here’s a brief introduction of this recognition method:

● The scanning unit scans the user’s lip movement picture

● Then it captures the first lip movement features according to the current lip movement picture

● According to the degree of matching between the first lip movement features and second lip moment features, it determines whether to unlock the device

As the first listed company in China’s AI data service industry, Nexdata is committed to breaking through technical bottlenecks and developed datasets for lip movement recognition, such as “1,998 People — Lip Language Video Data”, “156 Hours — Lip Sync Multimodal Video Data”, “1,066 People Living Face & Anti-Spoofing Data” etc. These own copyright datasets are collected with proper data collection authorization agreement.

1,998 People — Lip Language Video Data

Data Size: 1,998 people, at least 14 videos per person, 41,866 videos in total

Age Distribution: ranging from teenager to the elderly, the middle-aged and young people are the majorities

Collecting Environment: indoor scenes

Accuracy: the accuracy of annotation of shooting time and video content exceeds 95%

156 Hours — Lip Sync Multimodal Video Data

Recording Environment: Using quiet sunny room to stimulate daytime outdoor driving scenes, Signal to noise ratio 25~20dB

Recording Content: Short signals and spoken sentences

Recording Angle: Recording videos of front face, single side face, looking up, looking down, side face looking down and side face looking up all 6 different angles, and proximal and distant audio at the same time

Accuracy: Accuracy of sentence exceeds 95%

1,066 People Living Face & Anti-Spoofing Data

Data Size: 1,066 people, 235 videos and 63 images for each person

Collection Diversity: various postures, expressions, light condition, scenes, time periods and distances

Accuracy: the accuracy of actions exceeds 97%, the accuracy of naming the actions and lip language exceeds 97%

End

If you need data services, please feel free to contact us： [email protected]

The progress in the AI field cannot leave the credit of data. By improving the quality and diversity of datasets we can better unleash the potential of artificial intelligence, promote its applications of all walks of life. Only by continuously improving the data system, AI technology can better respond to the fast changing data requirements from market. If you have data requirements, please contact Nexdata.ai at [email protected].