Top insights for IT pros
From cybersecurity and big data to cloud computing, IT Brew covers the latest trends shaping business tech in our 4x weekly newsletter, virtual events with industry experts, and digital guides.
In Japan, Sony is using AI cameras that fit in the palm of your hand to analyze how customers interact with ads in 7-Eleven stores; 500 cameras have been deployed in convenience stores so far. The devices use “on-sensor AI processing to automatically detect when customers see digital signage, count how many people paused to view it, and measure the percentage of viewers,” Sony Semiconductors Solutions (SSS) announced in a press release last month.
IT Brew caught up with Eita Yanagisawa, the senior general manager of the system solutions business division at SSS, to chat 7-Eleven, edge AI, and protecting privacy.
This interview has been condensed for clarity and space; these responses come from an email interview.
What can you tell us about your latest endeavor with 7-Eleven and on-sensor AI processing?
Specifically, 7-Eleven uses the IMX500 and AITRIOS-based solution to understand the number of shoppers who view the in-store content and advertisements. This data could then be used by a back-end system to visualize patterns or purchasing behaviors of the customers who viewed the content, and how their purchase might be affected by the advertisement.
The sensor is the world’s first intelligent vision sensor with AI-processing capability.
What chips and processors are being used to power the tech?
For this solution in particular, with 7-Eleven and other convenience stores, we are utilizing the IMX500 sensor I mentioned to you before, which is powered by AITRIOS—SSS’s edge AI sensing platform. The IMX500 enables AI processing on the image sensor, the edge-most device, which serves as the entry point for image data.
How is Sony taking steps to protect consumers and mitigate any privacy risks when deploying these AI cameras in 7-Eleven or other places?
We enjoy answering this question, because it’s very unique to SSS and our IMX500 sensor. Traditional methods of edge AI that utilizes images are very difficult to handle. It creates extremely heavy data volume of images, and an expert is usually needed to extract the necessary information from the images.
What makes IMX500 different than other sensors is [that] it is an image sensor with AI processing capability. We can detect and output only the necessary metadata that is important to a business. This is in an effort to not only protect personally identifiable information, since it’s not needed for many solutions and insights, but also to reduce cloud costs with transmitting all visual data to the cloud before it can be processed.