Object Detection via Movement Vector Prediction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Natural user input devices, such as 3D cameras and eye tracking systems, suffer from imprecision in object selection due to sensor limitations and pointing inaccuracies, leading to prolonged adjustment processes.

Innovation Solution

A method that uses data points from natural user input devices to determine movement vectors, predict target areas, and assign interaction probabilities to objects, allowing for accurate selection without exact pointing precision, utilizing vector analysis and visual feedback to enhance selection efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If natural user input devices are used for object selection, then user interaction becomes more intuitive and accessible, but measurement precision and selection accuracy deteriorate due to sensor limitations and pointing inaccuracies

Engineering Contradiction:
Improveintuitiveness of user interactionVSAvoidobject selection accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The system performs preliminary actions by tracking user movement in advance and predicting the target object before the user completes the pointing action. The movement tracking module continuously monitors user input device position, and the target prediction module uses this historical movement data to anticipate which object the user intends to select, resolving the contradiction by providing accurate selection before precise pointing is achieved.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements feedback by providing visual indicators that show the predicted target object to the user during the selection process. This feedback loop allows the user to see what object the system predicts they want to select, enabling correction or confirmation of the prediction, thereby maintaining high selection accuracy despite imprecise pointing input.

Inventive Principle:
Principle #23Feedback

2Measurement precision

If visual feedback is provided to improve selection accuracy, then measurement precision increases, but loss of time increases due to prolonged adjustment processes

Engineering Contradiction:
Improveobject selection accuracyVSAvoidselection adjustment time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary target prediction based on tracked movement patterns before the user completes their pointing action. By analyzing the trajectory and velocity of the user input device, the system anticipates the target object in advance, providing visual feedback that confirms or corrects the prediction. This eliminates the need for prolonged adjustment processes while maintaining high selection accuracy.

Inventive Principle:
Principle #10Preliminary action

3Ease of operation

If the size of objects and spacing are increased to compensate for imprecision, then ease of operation improves, but area of stationary object increases which may not be desirable for all interfaces

Engineering Contradiction:
Improveselection easeVSAvoidobject display area
Core Design Contradiction:
Ease of operationVSArea of stationary object

Solution Approach 1:

The system replaces the mechanical approach of increasing object sizes with an intelligent prediction system. Instead of enlarging objects to make them easier to select, the system uses algorithms that track user movement patterns and predict the intended target. This substitution maintains the original object sizes and layout while achieving ease of operation through predictive accuracy.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP3015953B1Method and system for detecting objects of interest
Publication Date: 2019.12.18 4TIITOO
  • EP3015953B1 patent drawingFigure 1~2
  • EP3015953B1 patent drawingFigure 3~4
  • EP3015953B1 patent drawingFigure 5~6

AI summary

A method for detecting objects (28) a person has interest in interacting with comprises: receiving data points (38) of a natural user input device (26) adapted for tracking a movement of the person; determining at least one vector (40) from the data points (38) by grouping data points (38) into groups (42), wherein the data points (38) of a group (42) indicate a user movement in one direction; determining a map (44) of interaction objects (28), such that data points (38) are mapable to interaction objects (28); predicting a target area (48) from the at least one vector (40) by mapping the at least one determined vector (40) to the map (44) of interaction objects (28); and predicting at least one interaction object (28') at the target area (48) as interaction object of interest.