Search Apparatus Using Key Frame Pose and Time Intervals

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current techniques face challenges in improving search accuracy for moving images that exhibit human body movements similar to those in a query image, as they do not effectively utilize key frames and time intervals to match poses and movements.

Innovation Solution

A search apparatus and method that extracts key frames from a query moving image and searches for similar moving images based on the pose of a human body within these key frames and the time intervals between them, ensuring relevance and temporal similarity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If key frames are extracted and time intervals are considered for searching similar moving images, then search accuracy for movements is improved, but device complexity increases

Engineering Contradiction:
Improvesearch accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies segmentation by extracting key frames at specific time intervals from the query moving image rather than processing the entire video continuously. This divides the search process into discrete temporal segments, improving measurement precision for movement similarity while managing system complexity through selective frame processing

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements preliminary action by pre-extracting key frames and calculating their time intervals before performing the similarity search. This preparatory step establishes a temporal framework in advance, enabling more accurate movement comparison without adding complexity during the actual search execution

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If multiple key frames are extracted from query moving image, then pose similarity matching is improved, but loss of time increases

Engineering Contradiction:
Improvepose similarity matching accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies periodic action by extracting key frames at regular time intervals from the query moving image. This periodic sampling captures essential pose information across different time points, improving pose similarity matching accuracy while controlling processing time through systematic interval-based selection rather than continuous analysis

Inventive Principle:
Principle #19Periodic action

Data Source

PatentUS20250014342A1Search apparatus, search method, and non-transitory storage medium
Publication Date: 2025.01.09 NEC CORP
  • US20250014342A1 patent drawing
  • US20250014342A1 patent drawing
  • US20250014342A1 patent drawing

AI summary

A search apparatus (10) according to the present invention includes: a key frame extraction unit (11) that extracts a plurality of key frames from a query moving image; and a search unit (12) that searches for a moving image similar to the query moving image, based on a pose of a human body included in each of the plurality of key frames and a time interval between the plurality of key frames.