Object Tracking via Contour Filtering and Descriptor Assignment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing vision-based control systems for devices like gaming consoles and smartphones are inefficient due to high computational resource requirements, unsuitability for low light conditions, and need for specialized cameras, making them impractical for widespread use.

Innovation Solution

A computing device and method that filters contour lines and assigns descriptors to points in an image stream to reduce the dataset, allowing for effective object tracking without extensive computational resources, and enables tracking in low light conditions without specialized equipment, using a controller to determine relative distances and filter out non-descript points.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional vision-based control systems are used, then object tracking can be achieved, but computational resource requirements become excessively high

Engineering Contradiction:
Improveobject tracking accuracyVSAvoidcomputational resource requirements
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts and processes only the most critical information from image streams by detecting and tracking specific features (contour lines, edges, and key points) rather than processing the entire image data. This selective extraction of essential tracking elements significantly reduces computational resource requirements while maintaining adequate tracking accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the object detection process into distinct stages: contour line detection, feature point identification, descriptor assignment, and tracking. By dividing the complex image processing task into manageable segments, the system reduces overall computational complexity while preserving tracking precision through focused processing at each stage.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If traditional vision-based control systems are used, then object tracking can be achieved, but they require specialized cameras and good light conditions

Engineering Contradiction:
Improveobject tracking accuracyVSAvoidsuitability for low light conditions
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent employs standard, readily available cameras instead of requiring specialized expensive cameras. The system is designed to work with conventional image sensors that can capture sufficient data for tracking in various lighting conditions, making the solution more versatile and adaptable to different environments without requiring specialized hardware.

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

3Measurement precision

If traditional vision-based control systems are used, then object tracking can be achieved, but tracking speed becomes too slow to be effective

Engineering Contradiction:
Improveobject tracking accuracyVSAvoidtracking speed
Core Design Contradiction:
Measurement precisionVSSpeed

Solution Approach 1:

The patent performs preliminary actions by pre-processing images to detect contour lines and identify key features before the actual tracking begins. By preparing the image data in advance and establishing feature descriptors upfront, the system reduces processing time during real-time tracking, thereby increasing tracking speed while maintaining accuracy.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP2880510B1Improved video tracking
Publication Date: 2019.06.12 CRUNCHFISH
  • EP2880510B1 patent drawingFigure 1A~2
  • EP2880510B1 patent drawingFigure 3~5
  • EP2880510B1 patent drawingFigure 4B~4D

AI summary

A computing device for tracking an object in an image stream, said computing device comprising a memory and a controller, wherein said controller is configured to: receive an image stream comprising at least a first and a previous image of an object to be tracked, determine contour lines in at least said first image, wherein said contour lines comprises a plurality of points, determine and assign at least one descriptor to each point, filter out points based on the descriptors, determine relative distances between each point in said first picture with each point in said previous picture, which points have not been filtered out; and determine a maximum occurrence for the relative distances, wherein the maximum occurrence corresponds to a movement of said object to be tracked in the plane. A movement in a direction parallel to the line of sight is determined from an average position, a relative distance and slopes of linear fittings for each point-component of the points in the two images.