brighter AI
brighter AI
Next generation anonymization technology Deep Natural Anonymization Technology (DNAT)
Deep Natural Anonymization Technology (DNAT) anonymizes data without compromising its quality. It is possible to use AI while keeping attribute information that was lost in conventional anonymization such as mosaic.
Features of DNAT
- Re-recognition by face recognition technology is impossible
- DNAT synthesized faces are randomly generated and cannot be undone
- Attributes such as age and gender are retained for AI analysis and development
Use of DNAT
For example, when AI analyzes images from surveillance cameras, etc., it is only possible to simply confirm the number of people with conventional mosaic processing, but by using DNAT, it is possible to confirm the number of people by attribute (gender, age, wearing a mask, etc.). increase.
Results of machine learning analysis of Deep Natural Anonymization data and original data
Data: Cityscapes Validation Data Set
Machine Learning: Segmentation with Mask-RCNN
Inquiry/Document request
Macnica Brighter AI Responsible
- TEL：045-476-2010
- E-mail：brighterai@macnica.co.jp
Mon-Fri 8:45-17:30