| CPC G06Q 30/0631 (2013.01) [A45D 44/005 (2013.01); G06Q 30/0621 (2013.01); G06T 7/70 (2017.01); G06V 40/171 (2022.01); G06V 40/19 (2022.01); A45D 2044/007 (2013.01); G06T 2207/30201 (2013.01)] | 21 Claims |
|
1. A system for determining a skin tone of a user, the system comprising a server arrangement communicably coupled via a data communication network to a user device associated with the user, wherein the server arrangement is configured to:
obtain an input image capturing at least a face of the user;
analyze the input image to identify at least one first region of the input image that corresponds to a sclera of the user;
determine a representative pixel value of the sclera from pixel values of pixels in the at least one first region of the input image;
determine at least one second region of the input image that corresponds to a skin of the user, based on at least one of: an analysis of the input image, or an input received from the user device that is indicative of a selection by the user of the at least one second region of the input image;
detect and remove facial features that do not represent skin color from the at least one second region to define a skin sampling area;
normalize pixel values of pixels in the at least one second region or an entirety of the input image, based on a difference between the representative pixel value of the sclera and a reference pixel value to adjust for varied lighting conditions and environmental factors;
determine a representative pixel value of the skin sampling area from the normalized pixel values of the pixels in the at least one second region; and
select, from amongst a palette of skin tones, a skin tone whose pixel value matches with the representative pixel value of the skin sampling area, wherein the selected skin tone is determined as the skin tone of the user, the palette of skin tones being configured for accuracy based on an analysis of skin tones under various lighting conditions;
identifying identify, using machine learning, the first region of the input image utilizing a trained model on a predefined dataset of closeup images with pre-identified sclera regions;
select the at least one second region based on a size of the at least one second region, wherein the size of the at least one second region is greater than a predefined threshold size or is a largest among identified regions,
adjust variations in skin texture and color uniformity by utilizing tolerance levels for pixel value uniformity within selected regions ; and
determine the at least one second region based on the selection by the user via a graphical user interface for a personalized skin tone analysis.
|