Hello! I'm new to the field of photoplethysmography and I find it very interesting. I wanted to try it out with a monochrome camera that captures in the infrared range. I know that RGB holds more information and is more robust to artifacts, but this is just a simple test I want to perform.
My main question is ~ which BVP extraction methods are suited for single-channel analysis? Or, are any suited for this task? I understand that CHROM and POS rely on RGB, but would LGI or OMIT be possible?
My workflow would be as follows:
Perform ear mask segmentation and use it to calculate mean pixel intensity for each frame.
Process the 1D signal using either SVD or QR decomposition.
One confusion I have from the PVB methods.py file regards the structure of the signal array:
rPPG METHOD SIGNATURE
An rPPG method must accept theese parameters:
> signal -> RGB signal as float32 ndarray with shape [num_estimators, rgb_channels, num_frames], or a custom signal.
What does it mean by num_estimators? My raw signal would be of shape [ir_ch, num_frames], so basically a 1D array.
Hello! I'm new to the field of photoplethysmography and I find it very interesting. I wanted to try it out with a monochrome camera that captures in the infrared range. I know that RGB holds more information and is more robust to artifacts, but this is just a simple test I want to perform.
My main question is ~ which BVP extraction methods are suited for single-channel analysis? Or, are any suited for this task? I understand that CHROM and POS rely on RGB, but would LGI or OMIT be possible?
My workflow would be as follows:
One confusion I have from the PVB methods.py file regards the structure of the signal array:
What does it mean by
num_estimators? My raw signal would be of shape [ir_ch, num_frames], so basically a 1D array.Please let me know your thoughts!