Object Tracking Module for Video Frame Clustering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current mobile devices face challenges in efficiently tracking and detecting objects within video streams, particularly in providing an improved visual display and handling motion-based tracking and object detection effectively.

Innovation Solution

The implementation of an object tracking and detection module within mobile devices, which includes a motion tracker and an object detector, uses a combination of motion-based tracking and object detection methods to accurately pinpoint the location and size of objects within video frames, utilizing a memory buffer for feedback and optimizing computational resources.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If motion-based tracking and object detection are performed on all video frames, then tracking accuracy is improved, but computational load and energy consumption increase

Engineering Contradiction:
Improvetracking accuracyVSAvoidenergy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system performs full object detection only when necessary (when tracking confidence falls below threshold) rather than on every frame. Motion-based tracking is performed continuously as a lighter operation, while computationally intensive object detection is applied selectively to maintain accuracy without constant high energy consumption.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The system uses feedback from tracking confidence metrics to dynamically adjust the detection strategy. When tracking confidence is high, the system continues with motion-based tracking only. When confidence drops below a threshold, it triggers full object detection to correct potential drift, then returns to lighter tracking mode.

Inventive Principle:
Principle #23Feedback

2Measurement precision

If object detection is performed frequently, then detection accuracy is improved, but processing time increases

Engineering Contradiction:
Improvedetection accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system applies full object detection partially - only when tracking confidence indicates potential loss of target. Otherwise, it relies on continuous but lighter motion-based tracking, reducing overall processing time while maintaining detection accuracy when needed.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

Object detection is performed periodically based on confidence threshold violations rather than continuously. The system monitors tracking confidence over time and triggers detection only when necessary, creating a periodic rather than continuous detection rhythm that reduces processing time.

Inventive Principle:
Principle #19Periodic action

3Adaptability or versatility

If multiple objects are tracked simultaneously, then system versatility is improved, but device complexity increases

Engineering Contradiction:
Improvemulti-object tracking capabilityVSAvoidtracking system complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system segments the tracking task by maintaining separate tracking states and confidence metrics for each object. Each object is tracked independently with its own motion model and detection triggers, allowing multiple objects to be managed through modular, segmented processing rather than a monolithic complex system.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The same motion-based tracking and object detection mechanisms are applied universally to multiple objects. The system uses a unified approach for selecting and tracking different objects, reducing complexity by avoiding object-specific specialized code while maintaining versatility through generic multi-object handling.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3084577B1Selection and tracking of objects for display partitioning and clustering of video frames
Publication Date: 2018.08.29 QUALCOMM INC
  • EP3084577B1 patent drawingFigure 1
  • EP3084577B1 patent drawingFigure 2A
  • EP3084577B1 patent drawingFigure 2B

AI summary

A method includes selecting at least two objects within a primary video stream, generating a first video stream from the primary video stream that includes a first of the selected objects, and generating a second video stream from the primary video stream that includes a second of the selected objects. The primary video stream has a primary field of view, and the first and second video streams have respective first and second fields of view that are more narrow than the primary field of view. The first field of view includes a portion of the primary field of view that is not within the second field of view, and the second field of view includes a portion of the primary field of view that is not within the first field of view.