Marcus R Colbert

US Patent:

20180374249, Dec 27, 2018

Filed:

Jun 27, 2017

Appl. No.:

15/634171

Inventors:

Marcus C. COLBERT - Emerald Hills CA, US

Assignee:

Mad Street Den, Inc. - Fremont CA

International Classification:

G06T 11/60
G06K 9/62

Abstract:

Neural networks of suitable topology are trained with pairs of images, where one image of each pair depicts a garment, and the other image of each pair depicts the garment being worn by a model. Once trained, the neural network can synthesize an image based on a new image of a garment, where the synthesized image could plausibly have appeared in the training set, paired with the new image of the garment. Quantitative parameters controlling the image synthesis permit adjustment of features of the synthetic image, including the skin tone, body shape and pose of the model, accessories depicted in the synthetic image, and characteristics of the garment as depicted, such as length, sleeve style, collar style or tightness.

US Patent:

20170316576, Nov 2, 2017

Filed:

Jan 16, 2017

Appl. No.:

15/407069

Inventors:

Marcus C. COLBERT - Emerald Hills CA, US
Anand CHANDRASEKARAN - Chennai, IN

Assignee:

Mad Street Den, Inc. - Fremont CA

International Classification:

G06T 7/73
G06T 19/20
G06T 17/10
G06T 19/00

Abstract:

A subject's head position and motion can be tracked by analyzing a series of frames from a monocular camera and mapping distinguishing points visible in the frames onto an elliptical cylinder. The tracking data can be used to control a physical pan/tilt actuator or to reconfigure/reposition virtual objects, images of which can be synthesized and displayed, or composited back into the original frames and displayed.