Method and system for analyzing motion of subject using event-based sensor

Patent metadata

Field	Value
Publication number	US-9910059-B2
Application number	US-201414534199-A
Country	US
Kind code	B2
Filing date	Nov 6, 2014
Priority date	Jan 7, 2014
Publication date	Mar 6, 2018
Grant date	Mar 6, 2018

Abstract

A method is provided for analyzing motion of a subject. The method includes receiving multiple events from an event-based sensor, the events corresponding to a luminance change of the subject; calculating velocities of the events and directions of the events; generating a motion analysis table comprising an X axis classified into multiple velocity sections, a Y axis classified into multiple direction sections, and multiple cells representing frequencies of events corresponding to combinations of the velocity sections and the direction sections; and determining a motion of the subject based on the motion analysis table.

First claim

What is claimed is:

1. A method of determining motion of a subject, the method comprising:
generating a plurality of events by an event-based sensor in response to changes in luminance of the subject detected based on light received at pixels;
calculating velocities of the events and directions of the events based on correlating information from at least two pixels for each event;
generating a motion analysis table comprising an X axis classified into a plurality of velocity sections that represent different velocities, a Y axis classified into a plurality of direction sections that represent different directions, and a plurality of cells each representing a frequency of events corresponding to a different one of a plurality of combinations of the different velocities and the different directions; and
determining a motion of the subject based on a relative frequency of a combination of velocity and direction for a cell in comparison to other combinations for other cells in the motion analysis table.

2. The method of claim 1 , wherein determining the motion of the subject further comprises:
determining a motion direction of the subject as a representative direction of a direction section of a first cell, which has a highest frequency in the motion analysis table, based on a frame average velocity and the motion analysis table.

3. The method of claim 2 , wherein determining the motion direction of the subject as the representative direction of the direction section of the first cell based on the frame average velocity and the motion analysis table is performed when a frequency of the first cell is greater than a critical frequency of the motion analysis table, and the frequency of the first cell is greater than a product of a majority baseline ratio and one of a first frequency and a second frequency,
wherein the first frequency is a frequency of a second cell having a lowest frequency among cells corresponding to a velocity section of the first cell in the motion analysis table, and
wherein the second frequency is a frequency of a third cell having an opposite direction to a direction of the first cell among cells which correspond to the velocity section of the first cell in the motion analysis table.

4. The method of claim 1 , wherein determining the motion of the subject further comprises:
determining a motion velocity of the subject based on a frame average velocity and the motion analysis table.

5. The method of claim 4 , wherein determining the motion velocity of the subject based on the frame average velocity and the motion analysis table determines the motion velocity of the subject as a first normal motion velocity in a first case in which the frame average velocity is greater than a lower boundary velocity of a normal motion, a frequency of a first cell, which has a highest frequency in the motion analysis table, is greater than a first critical frequency of the motion analysis table, and the frequency of the first cell is greater than a product of a first majority baseline ratio and one of a first frequency and a second frequency,
wherein the first frequency is a frequency of a second cell having a lowest frequency among cells which correspond to a velocity section of the first cell in the motion analysis table, and
wherein the second frequency is a frequency of a third cell having an opposite direction to a direction of the first cell among cells which correspond to the velocity section of the first cell in the motion analysis table.

6. The method of claim 5 , wherein determining the motion velocity of the subject based on the frame average velocity and the motion analysis table determines the motion velocity of the subject as a second normal motion velocity in a second case, in which the frame average velocity is less than the lower boundary velocity of the normal motion, the frequency of the first cell is greater than a second critical frequency of the motion analysis table, and the frequency of the first cell is greater than a product of a second majority baseline ratio and one of the first frequency and the second frequency.

7. The method of claim 6 , wherein determining the motion velocity of the subject based on the frame average velocity and the motion analysis table determines the motion velocity of the subject to be 0 and that the subject is background in a case other than one of the first and second cases.

8. The method of claim 4 , wherein determining the motion velocity of the subject based on the frame average velocity and the motion analysis table comprises:
adjusting intervals of the velocity sections of the motion analysis table when a frame average velocity is less than a lower boundary velocity of a normal motion, a frequency of a first cell, which has a highest frequency in the motion analysis table, is greater than a critical frequency of the motion analysis table, and the frequency of the first cell is less than a product of a majority baseline ratio and one of a first frequency and a second frequency,
wherein the first frequency is a frequency of a second cell having a lowest frequency among cells which correspond to a velocity section of the first cell in the motion analysis table, and
wherein the second frequency is a frequency of a third cell having an opposite direction to a direction of the first cell among cells which correspond to the velocity section of the first cell in the motion analysis table.

9. The method of claim 8 , wherein adjusting the intervals of the velocity sections of the motion analysis table adjusts a portion of the intervals of the velocity sections of the motion analysis table.

10. The method of claim 8 , wherein adjusting the intervals of the velocity sections of the motion analysis table adjusts all of the intervals of the velocity sections of the motion analysis table.

11. The method of claim 8 , wherein generating the motion analysis table and determining the motion of the subject based on the motion analysis tables are performed again after adjusting the intervals of the velocity sections of the motion analysis table.

12. The method of claim 1 , wherein each of the events includes coordinate information, time information, and polarity information.

13. The method of claim 12 , wherein the polarity information is one of positive polarity information and negative polarity information.

14. The method of claim 12 , wherein calculating the velocities of the events and the directions of the events comprises:
calculating a velocity of a first event and a direction of the first event based on the first event and a second event, the first event being generated by an event sensing unit corresponding to a first pixel and the second event being generated by another event sensing unit corresponding to a second pixel located around the first pixel.

15. The method of claim 14 , wherein the second pixel is located within a window including the first pixel and N pixels located around the first pixel.

16. A subject motion determining system, comprising:
an event-based sensor configured to generate a plurality of events in response to changes in luminance of a subject detected based on light received at pixels;
a motion analysis table (MAT) generating unit configured to calculate velocities of the events and directions of the events based on correlating information from at least two pixels for each event, and to generate a motion analysis table comprising an X axis classified into a plurality of velocity sections that represent different velocities, a Y axis classified into a plurality of direction sections that represent different directions, an

Assignees

• Samsung Electronics Co Ltd

Inventors

• Bae Kwang Hyuk
• Kyung Kyu Min
• Kim Tae Chan
• Jeong Seong Yeong
• Cho Shung Han

Classifications

• G01P13/045 Primary

  with speed indication · CPC title