December 6, 2021
NTT e-Asia Corporation.
OCG Technology JSC
|\NTT e-Asia Corporation (established in Tokyo, President Hiroshi Tanabe (hereinafter, NTT e-Asia)) and OCG Technology JSC※1 (established in Hanoi, CEO Shunichi Kondo (hereinafter, OCG)) coordinated to work with Vietnamese engineers and started providing annotation service to create training data for Artificial Intelligence with large volumes with low cost.|
※1: A joint venture with Vietnam Posts and Telecommunications Group (VNPT), a state-owned telecommunications service provider in Vietnam.
1. Background and the purpose
In recent years, Artificial Intelligence (AI) has been widely applied in many fields and the demands are increasing dramatically. To make AI smarter, it is necessary to provide large training data for the AI to learn.
To create the training data, persons need to attach the meaning to the data, such as picture images, and this process is called Data Annotation.
For example, in order for the AI to learn the pens, the human specifies the position of the pen in the image and AI needs to understand this kind of shape is the pens. Since there are many different types of pens and the appearances are different depending on the angles, it is necessary to provide the different types of training data as many as possible.
To create the large training data, it needs a lot of manpower and this makes the costs high.
NTT e-Asia provides the service of creating the training data by annotation with low costs through its subsidiary in Vietnam, OCG, the IT company, together with high-skilled technicians in Vietnam. With this service, we can reduce 30% to 50% (Comparison at our companies) of the costs of creating the training data and this will contribute for spreading the AI more.
2. Service overview
Japanese companies that need the large training data for AI learning will provide raw image data for Annotation to NTT e-Asia and OCG will annotate utilizing the experienced Vietnamese operators in Hanoi and Japanese staffs are responsible for quality management.
As the example below, the staff will circle a rectangle on the object to distinguish the object, perform the zoning to divide the area for each object, then the staff will cross check the performance results and after final inspection by Japanese staffs, will deliver the formatted data as required to the customer.
3. Future plan
Not only in the annotation field, we cooperate with high-skilled engineers in Vietnam to deploy and develop AI engines with lower costs than in Japan.
Person in charge: Mr. Thang & Mr. Asano
OCG Technology Joint Stock Company
In charge: Mr.Yamamoto
★ There may be changes made to information above. Your kind understanding would be much appreciated.