Abstract: Embodiments described herein are directed to optical systems creating an optical image on an image sensor, the optical image having at least one foveated region of interest created using at least one localized magnification optical feature on an optical surface. The optical system can be useful to increase the number of infrared pixels that are used to image a specific target which is only in a part of the object scene, while still being able to image the whole scene in the RGB part of the spectrum. In a preferred embodiment, the optical system is used in an automotive scenario to image with higher resolution the driver in order to more efficiently run face tracking and recognition algorithms.

Abstract: A method for controlling an optical system including at least one active light-controlling element to better control the illumination of a scene when imaged on an image sensor is presented. This active light-controlling element is used to control the amount of light in an image zone depending on an electric signal from a controller. This allows imaging scenes with bright objects without over-exposure or scenes with dark objects without under-exposure on the image sensor. After image processing to reverse the effect of the active light-controlling element, the resulting processed images have an enhanced depth without the drawbacks associated to traditional HDR imaging. The method can use any shape of sub-elements in the active light-controlling element depending on the application. The method can also be used with multiple active light-controlling surfaces to further control the light on the image sensor, including separating the colors in the image.

Abstract: A method for training and using a convolutional neural network with images having on-purpose distortion is disclosed. By distorting on-purpose the image to create zone of interest with higher number of pixels, the resulting output from the neural network is improved. The distorted images are created using an imager device either comprising an optical system specifically designed to output distorted images or comprising software or hardware image distortion manipulation algorithm to create the distorted images from normal images. Also described is a method for training the neural network using a distorted image generator from various existing datasets.