“There is always safety in valor.” – Ralph Waldo Emerson
Across Asia, the European Union and the USA safety authorities are introducing regulations requiring systems to be integrated into new automobiles to enhance driver and pedestrian safety. The National Highway Traffic Safety Administration (NHTSA) estimates that in the US alone, drowsy driving leads to upwards of 100,000 police-reported crashes each year, resulting in a $12.5 billion cost to the economy. In the US, by May 2018 all new vehicles under 10,000 pounds must be equipped with rear-view cameras.
Since its inception, FotoNation has developed technologies that flawlessly detects human faces, eyes, irises, smiles blinks and head/eye movements. For more than 14 years we have delivered technologies noted by their high reliability, low power consumption, real-time processing capabilities, small footprint and scalability. Our unique expertise was gained in the high volume, highly demanding smartphone industry. As a result, FotoNation is ideally positioned to be a leading provider of the technologies that will power these new, mandated and intelligent automotive video and camera solution.
The advent of increasingly connected cars equipped with advanced safety features is creating the need to be able to more accurately detect, and for some applications even recognize drivers’ faces. Our expertise in face detection, recognition and tracking, accurate eye tracking as well as iris recognition, are ideally suited to enable the most advanced driver monitoring systems (DMS) in the market. Our systems deliver reliable results under all lighting conditions and on multiple types of driver facial images, even when partially obscured (e.g. when wearing sunglasses). Our solution can be deployed for a single- or dual- camera system to meet customer requirements.
FotoNation technology is uniquely designed to support applications such as Park Assist and 360 surround View. This can be accomplished by utilizing our robust algorithms in concert with Rear-view / Front view / Side-view cameras. FotoNation Geometrical Distortion Correction and Object Detection Engines are key IP cores that enable low power, low latency, high performance and maximum distance solutions, all of which is critical for object detection applications. Object detection enables pedestrian and other specific objects around the car to be recognized and so assist in driver decision making, which is a bridge into applications required to deliver semi-autonomous and fully-autonomous vehicles.
All of our solutions are designed with hardware-software hybrid implementation in mind. To achieve the highest performance and shortest execution times in as seamless way as possible FotoNation offers an in-house developed RTL IP Core Image Processing Unit (IPU).
A typical IPU configuration consumes less than 80 mW for multiple parallel applications. This low power image processing unit makes it possible to deliver long range object detection (faces at all angles, full and half body concurrently), coarse optical flow frame to frame registration, bi-cubic re-sampling or correction (lens + video image stabilization), 4k resolution @ 60 FPS, hybrid biometrics (Iris and Face Recognition) and support for many other computer vision and computational imaging features.