Image processing system and method
An image processing system for modifying one or more images of a user to change a cardiovascular activity that is measurable from the one or more images, the system comprising: an image receiving section configured to receive image data comprising a plurality of image frames; a region selection section configured to select one or more regions where the colour of the skin of the user changes temporally in the image data, wherein the colour changes are indicative of the cardiovascular activity of the user; and a modifying section configured to modify the colour change of the one or more selected regions in one or more image frames to create modified image data, as described elsewhere herein.
1. An image processing system for modifying one or more images of a user to change a measurable cardiovascular activity that is measurable from the one or more images, the system comprising:
an image receiving section configured to receive image data comprising a plurality of image frames taken of the user;
a region selection section configured to select one or more regions where temporal changes in a colour of the user's skin are present in the image data of the image frames, wherein the temporal changes in the colour are indicative of the cardiovascular activity of the user; and
a modifying section configured to modify the temporal changes in the colour of the selected one or more regions in one or more of the image frames to create modified image data related to the one or more images of the user such that the modified image data is not indicative of the cardiovascular activity of the user.
2. The image processing system of claim 1 , wherein the modifying section is configured to modify the temporal changes in the colour such that the modified image data is indicative of a predetermined cardiovascular activity representative of a cardiovascular activity.
3. The image processing system of claim 2 , wherein a temporal pattern of the temporal changes in the colour is selected by the user.
4. The image processing system of claim 1 , wherein the modifying section is configured to modify the temporal changes in the colour such that the modified image data is not indicative of cardiovascular activity.
5. The image processing system of claim 4 , in which the modifying section is configured to modify the temporal changes in the colour of the selected one or more regions by randomly perturbing a colour value within a predetermined range for each image frame in the plurality of image frames.
6. The image processing system of claim 1 , wherein the modifying section is configured to modify the temporal changes in the colour of each of the selected one or more regions to a rolling average of colour of each of the selected one or more regions over a predetermined number of preceding image frames.
7. The image processing system of claim 1 , wherein the modifying section is operable to apply a colour change generated by a machine learning model that is trained to control a measurement of the cardiovascular activity of the user, the measurement being generated by an imaging system configured to detect the cardiovascular activity of the user indicated in the image data.
8. The image processing system of claim 7 , wherein the machine learning model is an adversarial machine learning model.
9. The image processing system of claim 1 , wherein the region selection section is configured to select the one or more regions where the temporal changes in the colour of the skin of the user relative to a temporal colour change averaged across all of the skin of the user are present in the image data.
10. The image processing system of claim 1 , wherein the region selection section is configured to select the one or more regions where the temporal changes in the colour of the skin of the user are present over a predetermined number of frames.
11. A camera comprising the image processing system of claim 1 .
12. The image processing system of claim 1 , wherein the modifying section is configured to be activated in response to a predetermined trigger condition.
13. An image processing method for modifying one or more images of a user to change a measurable cardiovascular activity that is measurable from the one or more images, the method comprising the steps of:
receiving image data comprising a plurality of image frames taken of the user;
selecting one or more regions where temporal changes in a colour of the user's skin are present in the image data of the image frames, wherein the temporal changes in the colour are indicative of the cardiovascular activity of the user; and
modifying the temporal changes in the colour of the selected one or more regions in one or more of the image frames to create modified image data related to the one or more images of the user such that the modified image data is not indicative of the cardiovascular activity of the user.
14. A non-transitory machine-readable storage medium which stores computer software which, when executed by a computer, causes the computer to perform a method for modifying one or more images of a user to change a measurable cardiovascular activity that is measurable from the one or more images, the method comprising the steps of:
receiving image data comprising a plurality of image frames taken of the user;
selecting one or more regions where temporal changes in a colour of the user's skin are present in the image data of the image frames, wherein the temporal changes in the colour are indicative of the cardiovascular activity of the user; and
modifying the temporal changes in the colour of the selected one or more regions in one or more of the image frames to create modified image data related to the one or more images of the user such that the modified image data is not indicative of the cardiovascular activity of the user.