Face detection and blurring methods and systems
Methods, systems, and devices for generating a training set and using the training set to train an artificial intelligence algorithm to detect and blur faces in images of an interior space are described. An example method for generating a training set includes selecting a first image depicting an interior of a room, selecting a second image depicting at least a face of a person, generating, based on a transparency parameter and a roughness parameter, a semi-transparent surface image, selecting an illumination parameter to configure a first illumination level for the first image and a second illumination level for the second image, and generating a first composite image of the training data set by combining the illumination parameter, the first image, the second image, and the semi-transparent surface image.
1. A method of generating a training data set, comprising:
selecting a first image depicting an interior of a room;
selecting a second image depicting at least a face of a person;
generating, based on a transparency parameter and a roughness parameter, a semi-transparent surface image;
selecting an illumination parameter to configure a first illumination level for the first image and a second illumination level for the second image; and
generating a first composite image of the training data set by combining the illumination parameter, the first image, the second image, and the semi-transparent surface image.
2. The method of claim 1 , further comprising:
applying, based on the roughness parameter, a Gaussian blur kernel to the face of the person in the first composite image.
3. The method of claim 1 , further comprising:
generating a plurality of composite images of the training data set by varying values of the transparency parameter, the roughness parameter, and the illumination parameter for the first image, the second image, and the semi-transparent surface image.
4. The method of claim 3 , wherein generating the plurality of composite images comprises:
iterating over a range of values of a first parameter; and
generating a subset of the plurality of composite images for each value in the range of values of the first parameter, a constant value for a second parameter, and a constant value for a third parameter.
5. The method of claim 4 , wherein the first parameter is the transparency parameter, the second parameter is the roughness parameter, and the third parameter is the illumination parameter.
6. The method of claim 4 , wherein the first parameter is the roughness parameter, the second parameter is the transparency parameter, and the third parameter is the illumination parameter.
7. The method of claim 4 , wherein the first parameter is the illumination parameter, the second parameter is the roughness parameter, and the third parameter is the transparency parameter.
8. The method of claim 1 , further comprising:
training, based on the training data set, a deep neural network; and
using the deep neural network to blur at least one face in an inference image of the interior of the room.
9. A system for generating a training data set, comprising:
a processor and a memory including instructions stored thereupon, wherein the instructions upon execution by the processor cause the processor to:
select a first image depicting an interior of a room;
select a second image depicting at least a face of a person;
generate, based on a transparency parameter and a roughness parameter, a semi-transparent surface image;
select an illumination parameter to configure a first illumination level for the first image and a second illumination level for the second image; and
generate a first composite image of the training data set by combining the illumination parameter, the first image, the second image, and the semi-transparent surface image.
10. The system of claim 9 , wherein the processor is further configured to:
apply, based on the roughness parameter, a Gaussian blur kernel to the face of the person in the first composite image.
11. The system of claim 9 , wherein the processor is further configured to:
generate a plurality of composite images of the training data set by varying values of the transparency parameter, the roughness parameter, and the illumination parameter for the first image, the second image, and the semi-transparent surface image.
12. The system of claim 11 , wherein the processor is further configured, as part of generating the plurality of composite images, to:
iterate over a range of values of a first parameter; and
generate a subset of the plurality of composite images for each value in the range of values of the first parameter, a constant value for a second parameter, and a constant value for a third parameter.
13. The system of claim 12 , wherein the first parameter is the transparency parameter, the second parameter is the roughness parameter, and the third parameter is the illumination parameter.
14. The system of claim 12 , wherein the first parameter is the roughness parameter, the second parameter is the transparency parameter, and the third parameter is the illumination parameter.
15. The system of claim 12 , wherein the first parameter is the illumination parameter, the second parameter is the roughness parameter, and the third parameter is the transparency parameter.