中國的人臉識別算法在世界上處于領先地位,其中最好的算法能夠在不到一秒的時間內(nèi)識別出一千萬人而不會出現(xiàn)任何錯誤。
The top five of the 39 facial recognition algorithms in the world come from China, according to the Face Recognition Vendor Test (FRVT) conducted by the National Institute of Standards and Technology(NIST) under the US Department of Commerce, which released the test results for 2018 last Friday.
美國商務部國家標準與技術研究院上周五發(fā)布了2018年人臉識別測試的結果,根據(jù)該測試,在全球39種人臉識別算法中,排名前五的算法均來自中國。
YITU Technology, a Shanghai-based company that claimed both the top and second spot on the list, told the Global Times in a statement that its top algorithm can accurately recognize nearly every person in a sample base of 10 million.
總部位于上海的億圖信息科技公司在一則聲明中向《環(huán)球時報》透露,該公司的頂級算法可以在1000萬的樣本庫中準確識別出幾乎所有人。
In comparison, "it is very likely for a human brain to make a mistake in recognizing the identities of 100 people," the company said.
該公司稱,相比之下,“一個人的大腦很可能在識別100個人的身份時就犯一個錯誤。”
Algorithms from Beijing-based company Sense Time won the third and fourth spot, while the Chinese Academy of Sciences' Shenzhen Institutes of Advanced Technology (SIAT) claimed fifth in the test.
總部位于北京的商湯科技公司的算法獲得了第三名和第四名,而中國科學院深圳先進技術研究院則在測試中獲得了第五名。
YITU said that high facial recognition accuracy means that users gain better experiences when using related products.
億圖信息科技公司表示,面部識別的高精度意味著用戶在使用相關產(chǎn)品時會獲得更好的體驗。
For example, when a user uses facial ID for payment on a smartphone, the device will not authorize the payment for someone else with a similar appearance as the user, thecompany noted.
例如,該公司指出,當用戶在智能手機上使用面部識別進行支付時,該設備將不會授權與該用戶長相相似的其他人進行支付。