Subject stabilisation based on the precisely detected face position in the visual input and computer systems and computer-implemented methods for implementing thereof
First Claim
1. A computer-implemented method, comprising:
- obtaining, by at least one processor, a plurality of frames having a visual representation of a face of at least one person;
applying, by the at least one processor, for each frame, at least one multi-dimensional face detection regressor for fitting at least one meta-parameter to detect or to track a plurality of multi-dimensional landmarks that are representative of a presence of a face of at least one person in each respective frame;
separating, by the at least one processor, for each frame in the plurality of frames, the face of the at least one person from a background based on utilizing at least one deep learning algorithm;
applying, by the at least one processor, for each frame in the plurality of frames, at least one face movement detection algorithm to identify each displacement of each respective multi-dimensional landmark of the plurality of multi-dimensional landmarks between frames;
applying, by the at least one processor, for each two sequential frames in the plurality of frames, at least one face movement compensation algorithm that is configured to at least;
i) determine that a current displacement value of at least one respective multi-dimensional landmark of the plurality of multi-dimensional landmarks between two sequential frames exceeds a pre-determined threshold value, andii) re-draw the face of the at least one person for a particular frame of the two sequential frames, in which the current displacement value exceeds the pre-determined threshold value, to reduce the current displacement value of the at least one respective multi-dimensional landmark to an updated displacement value that is less than the pre-determined threshold value to generate a re-drawn face of the at least one person;
wherein the pre-determined threshold value is between 1 and 20 Hz; and
combining, by the at least one processor, the re-drawn face of the at least one person in the particular frame of the two sequential frames with the background to generate a face movement compensated output that stabilizes the visual representation of the face of the at least one person between the two sequential frames of the plurality of frames.
1 Assignment
0 Petitions
Accused Products
Abstract
In some embodiments, the present invention provides for an exemplary computer system that may include: a camera component configured to acquire a visual content, wherein the visual content having a plurality of frames with a visual representation of a face of a person; a processor configured to: apply, for each frame, a multi-dimensional face detection regressor for fitting at least one meta-parameter to detect or to track a plurality of multi-dimensional landmarks representative of a face; apply a face movement detection algorithm to identify each displacement of each respective multi-dimensional landmark between frames; and apply a face movement compensation algorithm to generate a face movement compensated output that stabilizes the visual representation of the face.
-
Citations
16 Claims
-
1. A computer-implemented method, comprising:
-
obtaining, by at least one processor, a plurality of frames having a visual representation of a face of at least one person; applying, by the at least one processor, for each frame, at least one multi-dimensional face detection regressor for fitting at least one meta-parameter to detect or to track a plurality of multi-dimensional landmarks that are representative of a presence of a face of at least one person in each respective frame; separating, by the at least one processor, for each frame in the plurality of frames, the face of the at least one person from a background based on utilizing at least one deep learning algorithm; applying, by the at least one processor, for each frame in the plurality of frames, at least one face movement detection algorithm to identify each displacement of each respective multi-dimensional landmark of the plurality of multi-dimensional landmarks between frames; applying, by the at least one processor, for each two sequential frames in the plurality of frames, at least one face movement compensation algorithm that is configured to at least; i) determine that a current displacement value of at least one respective multi-dimensional landmark of the plurality of multi-dimensional landmarks between two sequential frames exceeds a pre-determined threshold value, and ii) re-draw the face of the at least one person for a particular frame of the two sequential frames, in which the current displacement value exceeds the pre-determined threshold value, to reduce the current displacement value of the at least one respective multi-dimensional landmark to an updated displacement value that is less than the pre-determined threshold value to generate a re-drawn face of the at least one person; wherein the pre-determined threshold value is between 1 and 20 Hz; and combining, by the at least one processor, the re-drawn face of the at least one person in the particular frame of the two sequential frames with the background to generate a face movement compensated output that stabilizes the visual representation of the face of the at least one person between the two sequential frames of the plurality of frames. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
-
-
9. A system, comprising:
-
a camera component, wherein the camera component is configured to acquire a visual content, wherein the visual content comprises a plurality of frames having a visual representation of a face of at least one person; at least one processor; a non-transitory computer memory, storing a computer program that, when executed by the at least one processor, causes the at least one processor to; apply, for each frame of the plurality of frames, at least one multi-dimensional face detection regressor for fitting at least one meta-parameter to detect or to track a plurality of multi-dimensional landmarks that are representative of a presence of a face of at least one person in each respective frame; separate, for each frame in the plurality of frames, the face of the at least one person from a background based on utilizing at least one deep learning algorithm; apply, for each frame in the plurality of frames, at least one face movement detection algorithm to identify each displacement of each respective multi-dimensional landmark of the plurality of multi-dimensional landmarks between frames; apply, for each two sequential frames in the plurality of frames, at least one face movement compensation algorithm that is configured to at least; i) determine that a current displacement value of at least one respective multi-dimensional landmark of the plurality of multi-dimensional landmarks between two sequential frames exceeds a pre-determined threshold value, and ii) re-draw the face of the at least one person for a particular frame of the two sequential frames, in which the current displacement value exceeds the pre-determined threshold value, to reduce the current displacement value of the at least one respective multi-dimensional landmark to an updated displacement value that is less than the pre-determined threshold value to generate a re-drawn face of the at least one person; wherein the pre-determined threshold value is between 1 and 20 Hz; and combine the re-drawn face of the at least one person in the particular frame of the two sequential frames with the background to generate a face movement compensated output that stabilizes the visual representation of the face of the at least one person between the two sequential frames of the plurality of frames. - View Dependent Claims (10, 11, 12, 13, 14, 15, 16)
-
Specification