Object Recognition Device Using Prediction Information Extraction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing object recognition techniques using external sensors require processing all measured information to detect and recognize landmarks, which is inefficient and increases processing load.

Innovation Solution

An object recognition device and method that acquires external field information and object position data to extract prediction information, allowing for efficient recognition of objects by focusing on specific prediction information rather than all external field data, with optional attribute-based recognition methods.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If all external field information is processed to detect landmarks, then object recognition completeness is improved, but processing time and computational load increase

Engineering Contradiction:

Improveobject recognition completenessVSAvoidprocessing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary extraction of prediction information from external field data before full object recognition processing. By pre-identifying candidate regions or objects based on rough prediction criteria, the system prepares processed information in advance, allowing the main recognition algorithm to work only on predicted targets rather than all possible objects, thus reducing processing time while maintaining recognition completeness

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The invention extracts only the necessary prediction information from the complete external field data set. Instead of processing all sensor data through full recognition algorithms, the system extracts predicted object locations or characteristics as a subset of the total data, then applies detailed recognition only to these extracted candidates, achieving time efficiency without sacrificing recognition accuracy

Inventive Principle:

Principle #2Taking out (Extraction)

2Measurement precision

If all external field information is processed to detect landmarks, then object recognition accuracy is improved, but computational load increases

Engineering Contradiction:

Improveobject recognition accuracyVSAvoidcomputational load

Core Design Contradiction:

Measurement precisionVSPower

Solution Approach 1:

The object recognition process is divided into two segments: a prediction stage that quickly identifies candidate objects from external field information, and a recognition stage that applies accurate but computationally intensive algorithms only to these candidates. This segmentation allows high accuracy recognition to be achieved with reduced computational load compared to applying full recognition algorithms to all data

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs partial processing on the complete data set by extracting only prediction information that is sufficient for identifying candidate objects. This partial action (extracting predictions rather than full recognition features) reduces computational load while maintaining the accuracy needed for subsequent precise recognition of the identified objects

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS11215459B2Object recognition device, object recognition method and program

Publication Date: 2022.01.04 PIONEER IP

AI summary

The object recognition device acquires external field information from by an external detection device arranged on a movable body, and acquires object position information indicating a position of an object existing around the movable body. Then, the object recognition device extracts prediction information, predicted to include information indicating the object, from the external field information based on the object position information, and recognizes the object based on the prediction information.