Deep Learning with Energy-efficient Binary Gradient Cameras
This addresses power efficiency for embedded vision systems, offering a novel sensing approach with potential broad impact, though it is incremental in combining existing deep learning with new camera hardware.
The paper tackles the problem of high power consumption in embedded computer vision by using binary gradient cameras, showing that for tasks like object recognition and face detection, accuracy can match or exceed traditional images while reducing sensing power, and they demonstrate a prototype that recovers intensity information for human-in-the-loop applications.
Power consumption is a critical factor for the deployment of embedded computer vision systems. We explore the use of computational cameras that directly output binary gradient images to reduce the portion of the power consumption allocated to image sensing. We survey the accuracy of binary gradient cameras on a number of computer vision tasks using deep learning. These include object recognition, head pose regression, face detection, and gesture recognition. We show that, for certain applications, accuracy can be on par or even better than what can be achieved on traditional images. We are also the first to recover intensity information from binary spatial gradient images--useful for applications with a human observer in the loop, such as surveillance. Our results, which we validate with a prototype binary gradient camera, point to the potential of gradient-based computer vision systems.