Multi-resolution gesture recognition

• US 9,158,381 B2
• Filed: 02/25/2013
• Issued: 10/13/2015
• Est. Priority Date: 02/25/2013
• Status: Active Grant

• Current Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Original Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Inventors: Alaniz, Arthur
• Primary Examiner(s): Lee, Gene W
• Application Number:
  US13/776,131
• Publication Number:
  US 20140240213A1
• Time in Patent Office: 960 Days
• Field of Search: 345156-184
• US Class Current: 1/1
• CPC Class Codes: G06F 3/017 Gesture based interaction, e.g. based on a set of recognized hand gestures interaction based on gestures traced on a digitiser G06F3/04883

Abstract

Systems and methods for detecting and analyzing bodies visible to an input source are disclosed herein. Analysis of detected bodies can be used in gesture recognition. In particular embodiments, detection and recognition of hand and portions thereof, including fingers and portions thereof (e.g., fingertips) can be performed to recognize gestures related to the hand. Recognition and analysis of bodies visible to an input source can occur using, for example, blob detection techniques, machine learning techniques, and various calculative techniques including statistical analysis, application-specific equations, and estimations related to bodies anticipated to be visible to the input source.

Claims

1. A gesture recognition system, comprising:

a processor that executes the following computer executable components stored in a memory;

an image discovery component that records a motion of at least one body in a view of one or more sensors that collect input data that comprises the motion;

a blob identification component that identifies a first detail and a second detail of the at least one body, the blob identification component comprises a difference of Gaussians filter, wherein a large difference of Gaussian identifies the first detail and a small difference of Gaussian identifies the second detail; and

a gesture recognition component that identifies a gesture based on an analysis of the first detail, wherein the blob identification component masks the second detail from the gesture recognition component.

2. The system of claim 1, wherein the blob identification component comprises a median filter.

3. The system of claim 2, wherein the blob identification component comprises a band-pass filter.

4. The system of claim 1, wherein the blob identification component comprises a blob isolation component that isolates at least one portion of the body.

5. The system of claim 4, wherein the gesture recognition component identifies a gesture involving at least two portions of the body.

6. The system of claim 4, wherein the at least one portion of the body is a fingertip.