Tencent release 7 artificial intelligence cloud services

December 26, following the announcement of 11 overseas nodes open this month after the full, Tencent cloud once again announced in the field of artificial intelligence cloud service force, today officially to the global business to provide seven AI services, including face detection, facial features Face Recognition, Face Recognition, Image Labeling, ID OCR Recognition, Business Card OCR Recognition (OCR), Optical Character Recognition (OCR).

At the same time, Tencent announced the same period the industry’s largest AI free use of activities, involving a total of 8 AI services, that companies can enjoy the above-mentioned seven AI services 20,000 per month free use of concessions, but also get 60,000 per month Picture of Kam Kam Wong services, Tencent cloud picture Kam Wong has a precise distinction between pornography and sexy technology leadership.

If a company using all of the above services, a month to get 200,000 free concessions services, in accordance with the current growth trend of customers is expected to April 2017 activities ended, the company will send the whole Tencent value of at least 10 million AI Service, this preferential activity will carry on the artificial intelligence service to the small and medium-sized enterprise to have the obvious pull effect.

Wang Xinhui, vice president of Tencent cloud, said that this measure is designed to significantly reduce the use of artificial intelligence services cloud the threshold, the concept of artificial intelligence to implement, stimulate global business a new round of business innovation climax. After the release of the seven services, Tencent cloud will become the face recognition and image recognition products feature the most complete, free offers the greatest efforts of cloud service providers.

The release of the seven AI services are Tencent core technology capabilities, but also Tencent cloud and the industry’s top AI laboratory – Tencent excellent laboratory depth cooperation launched Tencent cloud “Vientiane excellent map” service, mainly attached to Vientiane excellent map Face recognition and image recognition of the two areas.

1, face detection: This is the first step in face recognition, can be used for any image, using a certain strategy to search to determine whether it contains a human face; if so, the feedback face position, size, gender And attitude and other attributes of the results, the accuracy rate can reach 99%;

2, facial features: can accurately locate the location of key areas of the face, including eyebrows, eyes, nose, mouth, face contours, the average error of less than 1.5 pixels, close to the level of manual labeling;

3, face comparison and verification: two people can calculate the similarity of face photographs to determine whether the same person. In the actual Internet financial scene test, self – photograph and ID thumbnails of the photo contrast is only 0.01% error rate, the rate of 95%;

4, face search: based on a given photo, and the database N face for comparison, to find out whether a sorting results for one person, to support the 10 million face database retrieval, the average retrieval time can reach milliseconds level;

5, the image tag: label covers hundreds of hot words, covering social life, people, scenery, architecture, common life objects and other categories, support different dimensions of the image semantic information extraction;

6, ID card OCR recognition: ID card to identify the positive and negative, a scan can identify all fields, the tilt, dark light, exposure, shadow and other abnormal conditions, the recognition of high accuracy. At the same time, through the adaptive judgment and correction technology, the recognition accuracy of ID on the ID card is improved to above 99.9%

7, business card OCR recognition: support name, phone number, QQ and other fields of automatic positioning and recognition, support 30 degrees within the angle of deflection, name field and cell phone number field of automatic recognition accuracy rate of 90%.

At the same time, Tencent cloud can also provide picture depth recognition engine service, based on user submitted image data for rapid learning, to provide customized image classification services, can be applied to food classification, flower classification, toy classification and other vertical field scenes .

Wang Huixing said, we see in practice, more and more the Internet industry, the financial industry and government customers to use artificial intelligence technology, artificial intelligence cloud service era has come, the future Tencent cloud will continue to be more comprehensive product services, Of the technical capacity, more preferential prices service enterprises.

The release of these seven services, but also to Tencent clouds of artificial intelligence cloud service layout has been further improved. The middle of this year, Tencent cloud release of large data service platform “a few wisdom strategy”, belongs to the field of artificial intelligence cloud services, a major layout. Tencent’s self-developed machine learning engine, as well as data extraction, processing analysis, report display, customer portraits, machine learning and other large data applications, the strategy provides open and common data application and analysis services, including user insight analysis, Flow analysis, cloud search, Wisdom natural language processing, intelligent voice services, wit machine learning, machine translation, general recommendation and other services.

The above services in Tencent WeChat, QQ, video, application of internal business such as treasure after a long period of practice and tempered. The number of intelligence in WeChat 16 billion daily level of multi-dimensional analysis of the scene, can be 6 seconds to return to the results; and application of the strategy in the use of several strategies recommended services, advertising download conversion rate increased by more than 49% of the game download and conversion The rate of increase of more than 38%.