System and method for producing custom fitted face masks and applications thereof
The present invention is generally directed to systems and methods for preparing and producing a custom-fitted face mask, including, as non-limiting examples, a CPAP nasal mask and a CPAP full face mask. At least one embodiment of the invention utilizes infrared (IR) lasers, such as, for example, those found on smartphone cameras, in order to generate a 3D point cloud model of an individual's face. This point cloud model is then used to produce a custom face mask cushion, which is used to customize a generic face mask to conform to the user's specific facial geometry.
1. A method for producing a custom-fitted face mask, the method comprising:
utilizing a mobile computing device to project a matrix of infrared wavelength laser points on a person's face, wherein each laser point in the matrix has a corresponding initial location on the person's face;
measuring one or more changes in the initial location of one or more of the infrared wavelength laser points, thereby producing one or more measurement change values;
generating a 3D point cloud model of the person's face using the one or more measurement change values; and
using the 3D point cloud model to manufacture a custom mask cushion that is capable of integrating with a generically-sized face mask, thereby producing a custom-fitted face mask for the person's face,
wherein the using the 3D point cloud model to manufacture a custom mask cushion comprises:
generating a uniform coordinate system for referencing equivalent points on a plurality of 3D face scans, each of the plurality of 3D face scans representing a different face;
fitting the generically-sized face mask on to a 3D face scan in the plurality of 3D face scans and quantifying an amount of mismatch between the generically-sized face mask and the 3D face scan;
clustering the plurality of 3D face scans to minimize an amount of overall mismatch that would result from using a fixed number of standard-sized face masks to cover all of the plurality of 3D face scans; and
generating a set of exemplar face scans based on the clustering.
2. The method of claim 1 , wherein the mobile computing device is a smartphone.
3. The method of claim 1 , wherein the matrix of infrared wavelength laser points is generated by a camera on the mobile computing device.
4. The method of claim 1 , wherein the measuring step and the generating step are both performed on the mobile computing device.
5. The method of claim 1 , wherein the custom mask cushion is manufactured by either 3D printing or injection molding.
6. The method of claim 1 , wherein the custom mask cushion is made of a soft, conformable material.
7. The method of claim 1 , wherein the generating a uniform coordinate system comprises warping a fixed face template to each 3D face scan in the plurality of 3D face scans.
8. The method of claim 7 , wherein the warping is achieved via solving a series of optimization problems in order to determine how best to warp the fixed face template to each 3D face scan in the plurality of 3D face scans.
9. The method of claim 8 , wherein solving the series of optimization problems comprises creating a series of 2D images corresponding to each of the plurality of 3D face scans by rendering each 3D face scan with chosen view angles and identifying facial landmarks on each series of 2D images.
10. The method of claim 9 , further comprising mapping the identified facial landmarks on each series of 2D images on to each corresponding 3D face scan in the plurality of 3D face scans.
11. The method of claim 10 , wherein the mapping the identified facial landmarks comprises using pixel coordinates of a camera on the mobile computing device.
12. The method of claim 7 , wherein warping the fixed face template to each 3D face scan in the plurality of 3D face scans further comprises applying a non-rigid transformation to the fixed face template, thereby permitting each vertex in the fixed face template to deform independently using each vertex's own affine transformation.
13. The method of claim 7 , further comprising:
generating a cushion template that defines a basic shape of the custom mask cushion;
positioning the cushion template on the fixed face template;
aligning each vertex on the cushion template with a nearest corresponding vertex on the fixed face template, thereby generating a combined face template; and
permitting the combined face template to deform to better fit the plurality of 3D face scans.
14. The method of claim 1 , wherein the clustering comprises utilizing a hierarchical clustering approach that identifies a face or a set of faces for which the generically-sized face mask would fit.