Japan's NEC has launched a new data analysis tool that uses AI in conjunction with face recognition.Called NeoFace Image Data Mining, the solution combines two technologies – face recognition and profiling across spatio-temporal data.”This enables people who have not been pre-registered to be identified almost immediately by their face from a large database of photographs”, President and CEO of NEC Corporation Takashi Niino, said at an event in Tokyo. “We have the number one technology in this area.”The product will be launched under the company's new flagship AI brand, NEC the Wise.”By 2020, there will be 53 billion things connected to the Internet and 7.7 billion people – that's 7 devices per person,” he said. “NEC is strong in visualisation and digitisation and there are three values we can offer to customers in the digital age – AI, connectivity and security.”In addition to searching for subjects based on facial image data, NeoFace Idm can also search for subjects who appear at specific times and locations, or who appear together with other specific individuals.The software can be utilized in a variety of applications, such as criminal investigations, searching for lost children, and providing improved customer service. When searching video where roughly one million individual instances of facial data appear, the software is capable of conducting searches within approximately 10 seconds.Spatio-Temporal Datatechnology creates groups of subjects based on their level of similarity and conducts high-speed searches for specific patterns.”In recent years, there is growing demand for advanced analysis of camera footage for utilization in security and marketing applications,” said Noritaka Taguma, General Manager, Transportation and City Infrastructure Division, NEC Corporation. “NeoFace Idm meets this demand by providing high-speed, high-precision searches for persons who appear in specific patterns, which could not be achieved through manual searches or conventional technology.”
NEC launches NeoFace Image Data Mining
