The new IMX500 and IMX501 intelligent vision sensors incorporate a logic processor and memory which allows cameras equipped with them to perform machine learning-powered computer vision tasks without the need for extra hardware. According to Sony, its new sensors could allow for the development of cheaper and more secure AI-equipped cameras.
In a press release announcing its new sensors, the company explained how they are able to process images using AI locally, saying:
“Signals acquired by the pixel chip are run through an ISP (Image Signal Processor) and AI processing is done in the process stage on the logic chip, and the extracted information is output as metadata, reducing the amount of data handled. Ensuring that image information is not output helps to reduce security risks and minimizing any privacy concerns.”
Intelligent vision sensors
Sony’s new AI image sensors have the ability to capture a regular 12-megapixel image and record 4K video at up to 60 frames per second. However, the sensors can also capture no image at all and just provide metadata based on what the sensor has seen.
Users will be able to write and rewrite their own AI models to the sensors’ embedded memory based on where the system is being used. For instance, Sony’s image sensors could be installed in cameras at the front of retail stores and can be used to count the number of visitors entering or even the number of products on the shelf much like the systems currently used in Amazon Go stores.
While there are plenty of positive applications for the company’s new image sensors, they could also be used for surveillance and monitoring if installed in CCTV cameras.
Sony has already started shipping test samples of its IMX500 image sensor to early customers and the company expects products that use the sensor to be available early next year.