The Team used ‘deep learning’ to develop a reconstruction algorithm called ‘FlatNet’ for lensless cameras which produces photorealistic images from lensless captures, reducing the gap in photograph quality between conventional cameras and ultra-thin lensless cameras