Log Data Generation for Image Recognition Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing techniques for information search in log data generated from image recognition processing, particularly in vehicle cameras, are inefficient due to increased data size, leading to slow search processing times.
Innovation Solution
A method of generating log data that includes detecting objects in each image frame, creating records with identification information and aggregated object counts, and adding metadata only when objects are detected, with the sequence of metadata determined by recognition frequency and detection score, and the data structure optimized to reduce unnecessary data storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If metadata is added to every record regardless of object detection results, then complete information is preserved, but data size increases and search processing time increases
Solution Approach 1:
The patent extracts metadata addition as a conditional operation rather than a universal one. Metadata is extracted and added only to records where objects are actually detected (aggregated number ≥ 1), while records with no detections (aggregated number = 0) exclude metadata. This selective extraction reduces data volume without losing information that would be relevant for search operations.
Solution Approach 2:
The patent applies local quality by making the data structure heterogeneous rather than uniform. Different records have different compositions based on their content: records with detections contain metadata while those without do not. This local differentiation optimizes each record individually, reducing overall data size while maintaining information quality where needed.
2Adaptability or versatility
If all object metadata is stored in every record, then comprehensive search capability is achieved, but data size expansion occurs
Solution Approach 1:
The patent extracts only the necessary metadata from the image recognition results and stores it conditionally. By checking the aggregated number of objects and only adding metadata when objects are present, the system extracts and stores only relevant information, reducing data quantity while maintaining search adaptability for cases where objects actually exist.
Solution Approach 2:
The patent applies partial action by not adding metadata to all records, but only to those where it is relevant (records with detected objects). This partial metadata inclusion is sufficient to maintain search capability for actual detections while avoiding the excessive data storage that would result from universal metadata addition.
3Productivity
If metadata sequence is arranged by recognition frequency and detection score, then search efficiency is improved, but additional processing complexity is introduced
Solution Approach 1:
The patent applies preliminary action by pre-sorting and arranging metadata in records according to recognition frequency and detection score before storage. This preliminary organization of data sequences enables more efficient search operations later, as the metadata is already optimized for retrieval based on importance and reliability metrics.
Data Source
AI summary
A method of generating log data that includes a plurality of records, in which information detected through image recognition processing on a plurality of image frames is recorded. The method includes, detecting objects in each image frame through the image recognition processing, generating a record including identification information of each image frame and the aggregated number of objects detected in the image frame for each image frame, and adding metadata relating to the detected objects to the corresponding record when the aggregated number of objects is equal to or greater than one. When the aggregated number of objects is zero, the metadata related to the detected objects is not added to the record.