Compressed Feature Descriptors for Efficient Image Matching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image processing technologies face inefficiencies in identifying features within images due to the large data requirements of feature descriptors, which hinder transmission, storage, and processing efficiency.

Innovation Solution

The method generates and compresses feature descriptors using gradient quantization and tree coding, allowing for efficient representation and comparison of features with a reduced number of bits, enabling feature identification without decompression.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If feature descriptors are represented with high precision using traditional methods, then feature identification accuracy is improved, but data transmission and storage efficiency deteriorates due to large number of bits required

Engineering Contradiction:

Improvefeature identification accuracyVSAvoidnumber of bits for feature descriptor

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The feature descriptor is segmented into multiple components: gradient orientation histograms are divided into multiple bins, and the histogram data is further segmented into significant and non-significant portions. This segmentation allows selective compression of less important data while preserving critical feature information, thereby reducing bit representation while maintaining identification accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The method extracts only the most significant gradient orientation information from the complete histogram data. By identifying and extracting dominant gradient directions and their magnitudes, the system discards redundant or less informative data, achieving compact representation that retains essential feature characteristics for accurate identification.

Inventive Principle:

Principle #2Taking out (Extraction)

2Quantity of substance

If feature descriptors are compressed to reduce bit representation, then transmission and storage efficiency is improved, but feature identification accuracy may deteriorate

Engineering Contradiction:

Improvenumber of bits for feature descriptorVSAvoidfeature identification accuracy

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The compression method applies different quality levels to different parts of the feature descriptor. Critical gradient orientation bins that contain significant feature information are preserved with high precision, while less important bins are compressed or discarded. This local quality differentiation ensures that the most discriminative features maintain their accuracy while reducing overall data size.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The system changes parameters such as the number of gradient orientation bins, the precision of magnitude representation, and the threshold for significant gradient selection. By dynamically adjusting these parameters based on feature importance and application requirements, the system optimizes the balance between compression ratio and identification accuracy.

Inventive Principle:

Principle #35Parameter changes

3Loss of information

If traditional feature descriptor methods are used, then feature representation completeness is maintained, but processing speed and system scalability deteriorate

Engineering Contradiction:

Improvefeature descriptor information completenessVSAvoidprocessing speed

Core Design Contradiction:

Loss of informationVSProductivity

Solution Approach 1:

The method performs preliminary compression and filtering of feature descriptors before they enter the main processing pipeline. By pre-identifying significant gradients and compressing the representation in advance, the system reduces the computational burden on subsequent processing stages, thereby improving overall processing speed without losing critical feature information.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system processes only the most relevant portions of the feature descriptor data rather than treating all data with equal detail. By focusing computational resources on significant gradient orientations and their neighborhoods, the system achieves faster processing while maintaining representation completeness for the most discriminative features.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentEP2359272B1Method and apparatus for representing and identifying feature descriptors utilizing a compressed histogram of gradients

Publication Date: 2023.07.26 NOKIA TECHNOLOGIES OY

Figure 1
Figure 2
Figure 3~4

AI summary

A method, apparatus and computer program product may be provided for generating a plurality of compressed feature descriptors that can be represented by a relatively small number of bits, thereby facilitating transmission and storage of the feature descriptors. A method, apparatus and computer program product may also be provided for permitting a compressed representation of a feature descriptor to be compared with a plurality of compressed representations of feature descriptors of respective predefined features. By permitting the comparison to be performed utilizing compressed representations of feature descriptors, a respective feature descriptor may be identified without having to first decompress the feature descriptor, thereby potentially increasing the efficiency with which feature descriptors may be identified.