US 12,094,087 B2
Systems, methods, and media for generating digital images using low bit depth image sensor data
Matthew Dutson, Madison, WI (US); and Mohit Gupta, Madison, WI (US)
Assigned to WISCONSIN ALUMNI RESEARCH FOUNDATION, Madison, WI (US)
Filed by WISCONSIN ALUMNI RESEARCH FOUNDATION, Madison, WI (US)
Filed on Mar. 4, 2022, as Appl. No. 17/687,390.
Prior Publication US 2023/0281770 A1, Sep. 7, 2023
Int. Cl. G06T 5/70 (2024.01); G06T 5/50 (2006.01); H04N 9/64 (2023.01); H01L 27/146 (2006.01); H01L 31/107 (2006.01)
CPC G06T 5/50 (2013.01) [G06T 5/70 (2024.01); H04N 9/646 (2013.01); G06T 2207/10016 (2013.01); G06T 2207/10024 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); H01L 27/14643 (2013.01); H01L 31/107 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system for generating range digital images, comprising:
an image sensor configured to generate low bit depth frames;
at least one processor that is programmed to:
receive, from the image sensor, a series of low bit depth frames;
provide low bit depth image information based on the series of low bit depth frames to a trained machine learning model, the trained machine learning model comprising:
a three-dimensional (3D) convolutional layer;
a two-dimensional (2D) convolutional long short term memory (LSTM) layer configured to receive an output of the 3D convolutional layer;
a concatenation layer configured to generate a tensor that includes a concatenation of an output of the 2D convolutional LSTM layer and the low bit depth image information; and
a 2D convolutional layer configured to generate an output based on the tensor generated by the concatenation layer; and
generate a high bit depth image of a scene based on an output of the 2D convolutional layer.