Cognate Image Sequence Identification Using Quality-Adaptive Indexing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in identifying cognate image sequences that have been altered by compression algorithms, transmission artifacts, and other changes, especially when operating downstream from the point of origin without access to the original sequences, requiring a method that can tolerate various artifacts and perform real-time identification with modest computational resources.
Innovation Solution
A system and method that uses an array of image indexes and quality indicators to compare incoming video streams with a database of known sequences, selecting appropriate image index thresholds based on quality indicators like average intensity, standard deviation, and colorfulness, to identify cognate sequences in real-time, even with compression and decompression artifacts, without significant local storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pixel by pixel comparison is used to match images, then identification accuracy is improved, but computational cost increases significantly
Solution Approach 1:
The patent divides the image into multiple blocks and computes separate indexes for each block rather than processing the entire image as a single unit. This segmentation approach reduces the computational complexity from O(N*M) where N is image size and M is database size, to a more manageable level by processing smaller blocks independently and combining results.
Solution Approach 2:
The patent extracts key features from images to create compact indexes (e.g., 10-100 bit binary strings) that represent the essential characteristics of images. This extraction process removes redundant information while preserving identification capability, enabling fast comparison with minimal computational resources.
2Productivity
If compression algorithms are applied to reduce image data, then transmission efficiency is improved, but image index accuracy deteriorates
Solution Approach 1:
The patent computes image indexes from the original uncompressed images before any compression or transmission occurs. These pre-computed indexes are stored and used for identification purposes, eliminating the need to reprocess compressed images and avoiding accuracy loss from compression artifacts.
Solution Approach 2:
The patent creates compact binary string copies (indexes) of the original image data that serve as surrogate representations. These indexes are transmitted and stored instead of full images, providing both compression benefits and maintaining identification accuracy by preserving key structural features in the binary representation.
3Speed
If real-time identification is implemented for high definition video streams, then response time is improved, but computational resource requirements increase
Solution Approach 1:
The patent processes video frames by dividing them into smaller blocks and computing indexes only for significant frames rather than all frames. This selective segmentation approach enables real-time processing of high definition streams by reducing the total number of comparisons while maintaining identification capability.
Solution Approach 2:
The patent performs partial processing by computing indexes only for significant frames that meet certain criteria (e.g., containing important visual information) rather than processing every frame. This partial action approach reduces computational load while maintaining sufficient identification accuracy for real-time applications.
4Adaptability or versatility
If downstream operation without original sequences is used, then system flexibility is improved, but identification reliability deteriorates
Solution Approach 1:
The patent pre-computes and stores indexes from original images at the point of origin before distribution. These pre-computed indexes serve as reliable reference data that can be used downstream without access to original sequences, maintaining identification reliability while enabling system flexibility and decentralized operation.
Data Source
AI summary
A system and method of identifying cognate image sequences is disclosed that examines significant frames of a stream of video images using an array of image indexes. The image index array includes image indexes obtained by at least two different image indexing methods. These are compared to a corresponding array of image indices of significant frames of known image sequences. An image quality indicator is used to determine which set of image index thresholds to use in making the comparison. These thresholds are more stringent for higher quality frames. Two image sequences are considered cognate when a string of sufficiently many sequential frame matches is established. In an alternate embodiment, image blurriness is also, or instead, used to determine the appropriate set of image index thresholds. The sets of image index thresholds are determined using machine learning on a curated set of representative images.