Method and apparatus for motion detection

• US 10,803,599 B2
• Filed: 09/26/2018
• Issued: 10/13/2020
• Est. Priority Date: 12/27/2012
• Status: Active Grant

• Current Assignee: Arria Data2text Limited
• Original Assignee: Arria Data2text Limited
• Inventors: Sripada, Gowri Somayajulu
• Primary Examiner(s): Bitar, Nancy
• Application Number:
  US16/142,445
• Publication Number:
  US 20190197697A1
• Time in Patent Office: 748 Days
• Field of Search: None
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30
• CPC Class Codes: G06T 2207/30192 Weather; Meteorology
• CPC Class Codes: G06T 7/20 Analysis of motion motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/43, H04N19/51
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06V 20/41 Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items segmenting video sequences G06V20/49

Abstract

Image analysis techniques may be employed to identify moving and/or static object within a sequence of spatial data frames (102, 300). Attributes of interest may be identified within a sequence of spatial data frames (102, 300). The attributes of interest may be clustered and examined across frames of the spatial data to detect motion vectors. A system (200) may derive information about these attributes of interest and their motion over time and identify moving and/or static objects, and the moving and/or static objects may be used to generate natural language messages describing the motion of the attributes of interest. Example uses include description of moving and/or static objects in data such as weather data, oil spills, cellular growth (e.g., tumor progression), atmospheric conditions (e.g., the size of a hole in the ozone layer), or any other implementation where it may be desirable to detect motion vectors in a sequence of spatial data frames.

Claims

1. An apparatus for identifying a moving object in spatial data, the apparatus comprising at least one processor and at least one memory storing instructions that, with the processor, cause the apparatus to:

identify a first cluster and a second cluster of a plurality of clusters in a sequence of spatial data frames, wherein the first cluster is associated with a first time value and the second cluster is associated with a second time value;

determine that the first cluster corresponds to the second cluster based on a similarity score computed for the first cluster at the first time value with the second cluster at the second cluster value, wherein the similarity score is based at least on a determined domain model;

wherein to determine that the first cluster corresponds to the second cluster, the apparatus is configured to;

select the first cluster and the second cluster for pairing in a circumstance where the similarity score is (1) a maximized cluster similarity score for the first cluster and the second cluster in the plurality of clusters, or (2) a maximized frame similarity score associated with the sequence of spatial data frames; and

generate, using a natural language generation system, an output text linguistically describing a moving object identified in part based on a motion vector between the first cluster and the second cluster.

2. The apparatus of claim 1, wherein the cluster similarity score is computed based in part on one or more of:

a determined a first time value location of each of the first cluster and the second cluster at the first time value; and

a determined second time value location of each of the first cluster and the second cluster at the second time value.

3. The apparatus of claim 1, wherein the sequence of spatial data frames represents spatial data of spatial-temporal data, received from a spatial data source, as a sequence of image-like objects.

4. The apparatus of claim 3, wherein the spatial-temporal data comprises a combination of time series and spatial data.

5. The apparatus of claim 1, wherein the sequence of spatial data frames defines one or more time value locations of the plurality of clusters at a plurality of time values.

6. The apparatus of claim 5, wherein the sequence of spatial data frames is represented as a set of frames, wherein each frame of the set of frames corresponds to a time value of the plurality of time values.

7. The apparatus of claim 1, wherein the first cluster is associated with a first set of attributes and the second cluster is associated with a second set of attributes.

8. The apparatus of claim 1, wherein the motion vector comprises a speed and a direction of motion.

9. The apparatus of claim 1, wherein the first cluster is determined as corresponding to the second cluster based on one of maximizing the cluster similarity score for the first cluster or maximizing a frame similarity score, the frame similarity score derived from the similarity score of each cluster at the first time value to at least one object at the second time value.

10. A method for identifying a moving object in spatial data, the method comprising:

identifying a first cluster and a second cluster of a plurality of clusters in a sequence of spatial data frames, wherein the first cluster is associated with a first time value and the second cluster is associated with a second time value;

determining that the first cluster corresponds to the second cluster based on a similarity score computed for the first cluster at the first time value with the second cluster at the second cluster value, wherein the similarity score is based at least on a determined domain model;

wherein determining that the first cluster corresponds to the second cluster comprises;

selecting the first cluster and the second cluster for pairing in a circumstance where the similarity score is (1) a maximized cluster similarity score for the first cluster and the second cluster in the plurality of clusters, or (2) a maximized frame similarity score associated with the sequence of spatial data frames; and

generating, using a natural language generation system, an output text linguistically describing a moving object identified in part based on a motion vector between the first cluster and the second cluster.

11. The method of claim 10, wherein the cluster similarity score is computed based in part on one or more of:

a determined a first time value location of each of the first cluster and the second cluster at the first time value; and

a determined second time value location of each of the first cluster and the second cluster at the second time value.

12. The method of claim 10, wherein the sequence of spatial data frames represents spatial data of spatial-temporal data, received from a spatial data source, as a sequence of image-like objects.

13. The method of claim 12, wherein the spatial-temporal data comprises a combination of time series and spatial data.

14. The method of claim 10, wherein the sequence of spatial data frames defines one or more time value locations of the plurality of clusters at a plurality of time values.

15. The method of claim 14, wherein the sequence of spatial data frames is represented as a set of frames, wherein each frame of the set of frames corresponds to a time value of the plurality of time values.

16. The method of claim 10, wherein the first cluster is associated with a first set of attributes and the second cluster is associated with a second set of attributes.

17. The method of claim 10, wherein the motion vector comprises a speed and a direction of motion.

18. The method of claim 10, wherein the first cluster is determined as corresponding to the second cluster based on one of maximizing the cluster similarity score for the first cluster or maximizing a frame similarity score, the frame similarity score derived from the similarity score of each cluster at the first time value to at least one object at the second time value.