US 11,903,730 B1
Body fat measurements from a two-dimensional image
Apoorv Chaudhri, Cupertino, CA (US); Siddhartha Chandra, Santa Clara, CA (US); Prakash Ramu, Portland, OR (US); Amit Kumar Agrawal, Santa Clara, CA (US); Sigal Raab, Sde Warburg (IL); Anantharanga Prithviraj, Santa Clara, CA (US); Ram Sever, Sunnyvale, CA (US); Ita Lifshitz, Kiryat Ono (IL); Ayush Sharma, Seattle, WA (US); Anna Shtengel, Santa Clara, CA (US); Gal Levi, Givataim (IL); and Rajesh Gautam, Cupertino, CA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Sep. 25, 2019, as Appl. No. 16/582,840.
Int. Cl. A61B 5/00 (2006.01); G06T 7/11 (2017.01); G06T 7/194 (2017.01); G06T 3/40 (2006.01); G06T 7/00 (2017.01); G06N 3/04 (2006.01); G06N 3/08 (2006.01); G16H 30/40 (2018.01); G06T 7/60 (2017.01)
CPC A61B 5/4872 (2013.01) [A61B 5/0077 (2013.01); A61B 5/7264 (2013.01); A61B 5/7278 (2013.01); A61B 5/742 (2013.01); G06N 3/04 (2013.01); G06N 3/08 (2013.01); G06T 3/40 (2013.01); G06T 7/0014 (2013.01); G06T 7/11 (2017.01); G06T 7/194 (2017.01); G06T 7/60 (2013.01); G16H 30/40 (2018.01); A61B 2560/0431 (2013.01); A61B 2576/00 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30196 (2013.01)] 22 Claims
OG exemplary drawing
 
1. A computer-implemented method, comprising:
receiving, from a portable device, a two-dimensional (“2D”) body image generated by a 2D camera of the portable device;
processing the 2D body image at a first bottleneck of a neural network to produce a first downsampled feature representation;
processing the first downsampled feature representation with a second bottleneck of the neural network to produce a second downsampled feature representation;
extracting first features from the second downsampled feature representation;
processing the second downsampled feature representation with a third bottleneck of the neural network to produce a third downsampled feature representation;
extracting second features from the third downsampled feature representation;
concatenating at least the first features, the second features, and an output of the neural network to produce concatenated features;
applying a linear function to the concatenated features to determine a body fat measurement representation;
determining a body fat measurement value from the body fat measurement representation; and
presenting, via a display of the portable device, the body fat measurement value.