Contactless image-based blood oxygen estimation
Systems, methods, apparatuses, and computer program products for contactless image-based blood oxygen estimation. A method may include receiving an image or video of a part of a subject captured by a camera of a computing device. The method may also include extracting a region of interest of the part of the subject from the image or video. The method may further include performing feature extraction of the region of interest. In addition, the method may include estimating a blood oxygen saturation level of the subject based on a spatial and temporal data analysis of more than two color channels. Feature extraction and estimation of the blood oxygen saturation level may include implementing a combination of spatial averaging, color channel mixing, and temporal trend analysis.
1 . A method, comprising:
receiving an image or video of a part of a subject captured by a camera of a computing device;
extracting a region of interest of the part of the subject from the image or video;
performing feature extraction of the region of interest;
estimating a blood oxygen saturation level of the subject based on a spatial and temporal data analysis of more than two color channels, wherein feature extraction and estimation of the blood oxygen saturation level comprises implementing a combination of spatial averaging, color channel mixing, and temporal trend analysis;
extracting a remote photoplethysmogram signal from the spatial averaging;
determining a heart rate of the subject from the photoplethysmogram signal;
calculating, based on the heart rate, DC and AC components of red, green, and blue color channels of the camera based on the spatial averaging; and
generating a regression model from pairwise ratios of the spatial averaging.
2 . The method according to claim 1 , wherein estimation of the blood oxygen saturation level comprises implementing a multi-channel ratio-of-ratios feature vector with a narrow adaptive bandpass filter.
3 . The method according to claim 1 , wherein estimation of the blood oxygen saturation level comprises implementing a neural network.
4 . The method according to claim 3 ,
wherein the neural network is a convolutional neural network, and
wherein estimating the blood oxygen saturation level comprises feeding red, green, and blue time series data of the spatial averaging into a convolutional neural network.
5 . The method according to claim 3 ,
wherein the neural network comprises a structure with at least one of feature extraction from skin color signals of the subject, and channel mixing of red, green, and blue color channels of the camera.
6 . The method according to claim 1 , further comprising:
estimating a weight vector from the regression model,
wherein the blood oxygen saturation level is estimated based on the estimated weight vector.
7 . An apparatus, comprising:
at least one processor; and
at least one memory comprising computer program code,
the at least one memory and the computer program code are configured, with the at least one processor, to cause the apparatus at least to receive an image or video of a part of a subject captured by a camera of a computing device;
extract a region of interest of the part of the subject from the image or video;
perform feature extraction of the region of interest;
estimate a blood oxygen saturation level of the subject based on a spatial and temporal data analysis of more than two color channels, wherein feature extraction and estimation of the blood oxygen saturation level comprises implementing a combination of spatial averaging, color channel mixing, and temporal trend analysis;
extract a remote photoplethysmogram signal from the spatial averaging;
determine a heart rate of the subject from the photoplethysmogram signal;
calculate, based on the heart rate, DC and AC components of red, green, and blue color channels of the camera based on the spatial averaging; and
generate a regression model from pairwise ratios of the spatial averaging.
8 . The apparatus according to claim 7 , wherein estimation of the blood oxygen saturation level comprises implementing a multi-channel ratio-of-ratios feature vector with a narrow adaptive bandpass filter.
9 . The Apparatus according to claim 7 , wherein estimation of the blood oxygen saturation level comprises implementing a neural network.
10 . The apparatus according to claim 9 ,
wherein the neural network is a convolutional neural network, and
wherein estimating the blood oxygen saturation level comprises feeding red, green, and blue time series data of the spatial averaging into a convolutional neural network.
11 . The apparatus according to claim 9 , wherein the neural network comprises a structure with at least one of feature extraction from skin color signals of the subject, and channel mixing of red, green, and blue color channels of the camera.
12 . The apparatus according to claim 7 , wherein the at least one memory and the computer program code are further configured, with the at least one processor, to cause the apparatus at least to:
estimate a weight vector from the regression model,
wherein the blood oxygen saturation level is estimated based on the estimated weight vector.
13 . A computer program, embodied on a non-transitory computer readable medium, the computer program comprising computer executable code, which, when executed by a processor, causes the processor to:
receive an image or video of a part of a subject captured by a camera of a computing device;
extract a region of interest of the part of the subject from the image or video;
perform feature extraction of the region of interest;
estimate a blood oxygen saturation level of the subject based on a spatial and temporal data analysis of more than two color channels, wherein feature extraction and estimation of the blood oxygen saturation level comprises implementing a combination of spatial averaging, color channel mixing, and temporal trend analysis;
extract a remote photoplethysmogram signal from the spatial averaging;
determine a heart rate of the subject from the photoplethysmogram signal;
calculate, based on the heart rate, DC and AC components of red, green, and blue color channels of the camera based on the spatial averaging; and
generate a regression model from pairwise ratios of the spatial averaging.
14 . The computer program according to claim 13 , wherein estimation of the blood oxygen saturation level comprises implementing a multi-channel ratio-of-ratios feature vector with a narrow adaptive bandpass filter.
15 . The computer program according to claim 13 , wherein estimation of the blood oxygen saturation level comprises implementing a neural network.
16 . The computer program according to claim 15 ,
wherein the neural network is a convolutional neural network, and
wherein estimating the blood oxygen saturation level comprises feeding red, green, and blue time series data of the spatial averaging into a convolutional neural network.
17 . The computer program according to claim 15 , wherein the neural network comprises a structure with at least one of feature extraction from skin color signals of the subject, and channel mixing of red, green, and blue color channels of the camera.
18 . The computer program according to claim 13 , wherein the processor is further caused to:
estimate a weight vector from the regression model,
wherein the blood oxygen saturation level is estimated based on the estimated weight vector.