Image processing and presentation techniques for enhanced proctoring sessions
Systems and methods for providing remotely proctored at-home health testing and diagnostics are provided herein. In particular, systems and methods for improving proctor performance by increasing accuracy and reducing proctor time are disclosed. In some embodiments, images and/or video of a patient may be enhanced to aid the proctor. In some embodiments, proctor performance may be evaluated. In some embodiments, proctor time per patient may be reduced by automating one or more steps in a proctored testing process.
1. A computer-implemented method for a remote diagnostic testing platform, the computer-implemented method comprising:
receiving, by a computing system, a video feed from a user device of a user engaging in an on-demand test session;
analyzing, by the computing system, the video feed to automatically determine that a particular step in the on-demand test session has been reached by the user;
based on detection of the particular step, storing, by the computing system, a plurality of subsequently-received image frames of the video feed to a first buffer;
evaluating, by the computing system, the plurality of image frames stored in the first buffer against a set of criteria, wherein the evaluation comprises evaluating motion blur, a difference between a frame of the plurality of image frames and an adjacent image frame of the plurality of image frames, and a closeness of a camera exposure level to a predetermined exposure level;
selecting, by the computing system, a subset of the image frames stored in the first buffer based at least in part on the evaluation;
storing, by the computing system, the selected subset of image frames to a second buffer;
processing, by the computing system, the subset of image frames stored in the second buffer to generate a composite image; and
performing, by the computing system, one or more operations using the composite image.
2. The computer-implemented method of claim 1 , further comprising detecting, by the computing system, that the particular step in the on-demand test session has ended.
3. The computer-implemented method of claim 1 , wherein performing comprises presenting the composite image to a proctor.
4. The computer-implemented method of claim 1 , wherein performing comprises creating or modifying a training data set, the training data set to be used for training a machine learning model.
5. The computer-implemented method of claim 1 , wherein performing comprises:
extracting, by the computing system, information from the composite image; and
querying, by the computing system, a database using the extracted information.
6. The computer-implemented method of claim 1 , wherein processing comprises:
extracting, by the computing system, a region of interest from each image frame stored in the second buffer;
using, by the computing system, template-matching to overlay image data extracted from the frames stored in the second buffer;
processing, by the computing system, the overlaid image data to enhance the image; and
combining, by the computing system, the processed overlaid image data to form a composite image.
7. The computer-implemented method of claim 6 , wherein enhancing the image comprises any combination of one or more of suppressing noise, normalizing illumination, rejecting motion blur, enhancing resolution, rotating, keystone correcting, or increasing image size.
8. The computer-implemented method of claim 7 , wherein normalizing illumination comprises:
determining, by the computing system, a size of the region of interest in at least one dimension;
accessing, by the computing system, a kernel for normalizing illumination levels in images;
dynamically adjusting, by the computing system, a size of the kernel in at least one dimension based at least in part on the determined size of the region of interest; and
applying, by the computing system, the adjusted kernel to one or more patches of the region of interest to normalize one or more levels of illumination within the region of interest.
9. The computer-implemented method of claim 6 , further comprising:
providing, to a proctor computing device, the composite image.