Back to results
US 11,113,815 B2Granted
Video processing methods and software architectures for analyzing transformation in objects
AssigneePurdue Research Foundation
InventorsPavlos P Vlachos, Brett Meyers
Life of this patent
Estimating the term…
Abstract
Video processing methods and the associated system architecture for measuring transformation in objects, including pupils, entail the following steps: 1. Motion correction; 2. Object (eye) detection; 3. Image correction; and 4. Fourier-based analysis for item (in some embodiments the item is a pupil) motion estimation.
Claims20
A software architecture encoded on a non-transitory computer readable medium, the software architecture processing a method of processing a video, wherein the method comprises:
- discarding a set of unusable frames from a video, wherein the video comprises an object of interest;
- selecting a set of usable frames and a reference frame from the video;
- performing a first registration of each usable frame of the set of useable frames to the reference frame, thereby generating a set of updated usable frames;
- storing a plurality of variables corresponding to the object of interest from each updated usable frame using a feature extraction algorithm;
- detecting the object of interest from the each updated usable frame using the feature extraction algorithm;
- determining a sub region containing the object of interest in the each updated useable frame;
- performing color image contrast enhancement on each sub region of the each updated usable frame, thereby producing a plurality of enhanced updated usable sub region frames;
- converting each enhanced updated useable sub region frame of the plurality of enhanced updated usable sub region frames into a grayscale sub region frame, thereby producing a plurality of greyscale sub region frames;
- performing a second registration of each greyscale sub region frame of the plurality of greyscale sub region frames, thereby generating a plurality of updated greyscale sub region frames;
- and
- calculating transformation parameters between the plurality of updated greyscale sub region frames.
8 more elements
A software architecture encoded on a non-transitory computer readable medium, the software architecture processing a method of processing a video, wherein the method comprises:
- discarding a set of unusable frames from a video, wherein the video comprises an object of interest;
- selecting a set of usable frames and a reference frame from the video;
- performing a first registration of each usable frame of the set of useable frames to the reference frame, thereby generating a set of updated usable frames;
- storing a plurality of variables corresponding to the object of interest from each updated usable frame using a feature extraction algorithm;
- detecting the object of interest from the each updated usable frame using the feature extraction algorithm;
- determining a sub region containing the object of interest in the each updated useable frame;
- performing color image contrast enhancement on each sub region of the each updated usable frame, thereby producing a plurality of enhanced updated usable sub region frames;
- converting each enhanced updated useable sub region frame of the plurality of enhanced updated usable sub region frames into a grayscale sub region frame, thereby producing a plurality of greyscale sub region frames;
- performing a second registration of each greyscale sub region frame of the plurality of greyscale sub region frames, thereby generating a plurality of updated greyscale sub region frames;
- calculating transformation parameters between the plurality of updated greyscale sub region frames;
- and
- calculating a scaling coefficient between the plurality of updated greyscale sub region frames.
9 more elements
A software architecture encoded on a non-transitory computer readable medium, the software architecture processing a method of processing a video, wherein the method comprises:
- discarding a set of unusable frames from a video, wherein the video comprises an object of interest, wherein the object of interest comprises an eye;
- selecting a set of usable frames and a reference frame from the video;
- performing a first registration of each usable frame of the set of useable frames to the reference frame, thereby generating a set of updated usable frames;
- storing a plurality of variables corresponding to the object of interest from each updated usable frame using a feature extraction algorithm;
- detecting the object of interest from the each updated usable frame using the feature extraction algorithm;
- determining a sub region containing the object of interest in the each updated useable frame;
- performing color image contrast enhancement on each sub region of the each updated usable frame, thereby producing a plurality of enhanced updated usable sub region frames;
- converting each enhanced updated useable sub region frame of the plurality of enhanced updated usable sub region frames into a grayscale sub region frame, thereby producing a plurality of greyscale sub region frames;
- performing a second registration of each greyscale sub region frame of the plurality of greyscale sub region frames, thereby generating a plurality of updated greyscale sub region frames;
- calculating transformation parameters between the plurality of updated greyscale sub region frames;
- and
- calculating a scaling coefficient between the plurality of updated greyscale sub region frames.
9 more elements
Dependencies are inferred from the claim text.