Duplicate Scene Detection for Edge AI Image Workflows
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Edge devices generate duplicative image data of the same scene, leading to exponential data growth and inefficient resource allocation in centralized processing, as each instance creates separate workflows and storage requirements.
Innovation Solution
A storage system detects duplicate scenes using inference labels and contextual information, implementing policies to manage, migrate, or delete duplicate data, thereby optimizing storage and workflows.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If edge devices send all image data to centralized processing, then complete data analysis is achieved, but data storage costs and processing time increase exponentially due to duplicate scenes
Solution Approach 1:
The system performs preliminary duplicate detection and filtering at the edge device before transmitting image data to the centralized system. By identifying and eliminating duplicate scenes in advance using inference labels and contextual information, the system reduces the volume of data requiring centralized processing while ensuring that unique, valuable data is transmitted for complete analysis.
2Measurement precision
If all image data is processed centrally, then comprehensive scene analysis is achieved, but response time deteriorates due to large data volumes
Solution Approach 1:
Duplicate detection and filtering is performed in advance at the edge device before centralized processing. This preliminary action reduces the data volume that needs to be transmitted and processed centrally, thereby maintaining comprehensive scene analysis capability while significantly reducing processing response time.
Solution Approach 2:
The system divides the processing workflow into two segments: edge-based duplicate detection using inference labels and contextual information, and centralized processing of only unique scenes. This segmentation allows each component to focus on specific tasks, improving overall efficiency and response time.
3Reliability
If duplicate images are stored separately, then data integrity is maintained, but resource allocation efficiency decreases
Solution Approach 1:
The system performs preliminary duplicate detection using inference labels and contextual information before data is fully ingested into the centralized system. By identifying duplicates in advance, the system can consolidate storage requirements and optimize resource allocation while maintaining data integrity through selective retention of unique scenes.
Data Source
AI summary
Provided is a computer-implemented method, system, and computer program product for detecting and processing a duplicate scene. A processor may receive image data from a plurality of images of a scene into a storage system, the image data including inference labels and contextual information associated with each of the plurality of images. The processor may detect, based on the inference labels and contextual information, that one or more images of the plurality of images is a duplicate of the scene. The processor may execute, in response to detecting that the one or more images of the plurality of images is the duplicate of the scene, an action related to the plurality of images.