Image Processing Method for AR Object Recognition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current image processing methods for augmented reality in electronic devices face challenges with high network bandwidth consumption when transmitting continuous image frames for object recognition, and they also burden the device with continuous image processing tasks.

Innovation Solution

An image processing method that detects motion between the electronic device and objects, allowing for the recognition of objects by determining if the same object exists in continuous frames, thereby reducing the need for continuous image transmission and processing by using a feature point extraction and compression module to transmit only necessary data to a server for comparison.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If continuous image frames are transmitted to a server for object recognition, then object recognition accuracy is improved, but network bandwidth consumption increases

Engineering Contradiction:
Improveobject recognition accuracyVSAvoidnetwork bandwidth consumption
Core Design Contradiction:
Measurement precisionVSLoss of energy

Solution Approach 1:

The patent extracts only the essential feature points from continuous image frames instead of transmitting entire frames. The feature point extraction module identifies and transmits only the critical visual features necessary for object recognition, significantly reducing network bandwidth consumption while maintaining recognition accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies partial action by transmitting only a subset of image data (feature points) rather than complete image frames. This selective transmission approach provides sufficient information for object recognition without the excessive bandwidth consumption of full-frame transmission.

Inventive Principle:
Principle #16Partial or excessive action

2Measurement precision

If the electronic device directly performs continuous image frame processing, then object recognition capability is improved, but device processing burden increases

Engineering Contradiction:
Improveobject recognition capabilityVSAvoidprocessing burden
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the image processing task into two parts: local feature point extraction performed by the electronic device, and comprehensive object recognition performed by the server. This division reduces the processing burden on the device while maintaining recognition capability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The feature point extraction module acts as an intermediary that prepares simplified image data before transmission. By preprocessing images to extract only essential features, the device reduces its processing burden while ensuring the server receives sufficient information for accurate recognition.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Loss of energy

If feature point data is transmitted to the server, then network bandwidth usage is reduced, but data transmission requirements change

Engineering Contradiction:
Improvenetwork bandwidth usageVSAvoiddata transmission requirements
Core Design Contradiction:
Loss of energyVSDevice complexity

Solution Approach 1:

The patent changes the parameters of transmitted data from complete image frames to extracted feature points. This parameter transformation significantly reduces data volume and network bandwidth usage, though it requires implementing feature point extraction functionality in the device.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP2950273B1Image processing method and electronic device implementing the same
Publication Date: 2022.05.04 SAMSUNG ELECTRONICS CO LTD
  • EP2950273B1 patent drawingFigure 1
  • EP2950273B1 patent drawingFigure 2
  • EP2950273B1 patent drawingFigure 3

AI summary

An image processing method and an electronic device implementing the method are provided. The method includes the electronic device sequentially receiving an image captured by a camera and determining whether a frame of the received image satisfies a predetermined condition. If the predetermined condition is satisfied, the electronic device recognizes an object in the frame. Then the electronic device tracks the object in the frame through tracking data created based on a feature extracted from the recognized object.