Image Segmentation via Partial Distance Pruning and Lookup Tables

Solution Overview

Problem

Conventional image segmentation techniques for embedded systems require significant computational resources and circuit size due to the high amount of calculation needed to determine the distance between input data and representative data, making it challenging to efficiently cluster high-resolution image data.

Innovation Solution

An image processing method that calculates partial distances between a pixel of interest and reference pixels, omits the calculation of total distance if the partial distance exceeds the shortest total distance, and categorizes the pixel based on the nearest reference pixel, thereby reducing unnecessary calculations and speeding up the clustering process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional clustering algorithms calculate the distance between input data and each representative data to segment images, then segmentation accuracy is maintained, but the computational load and circuit size increase significantly

Engineering Contradiction:

Improvesegmentation accuracyVSAvoidcircuit size

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by pre-calculating and storing distance values in a lookup table before the actual clustering process. The distance between pixels and representative data points is computed in advance and stored, so that during segmentation, only table lookup operations are needed instead of performing complex distance calculations, thereby reducing circuit complexity while maintaining accuracy

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent segments the computational process into two distinct phases: an offline preprocessing phase where distance values are calculated and stored in a lookup table, and an online execution phase where only simple table lookups and comparisons are performed. This segmentation allows the complex calculation work to be done when computational resources are abundant, while the actual real-time segmentation uses minimal computational resources

Inventive Principle:

Principle #1Segmentation

2Measurement precision

If conventional clustering algorithms calculate the distance between input data and each representative data to segment images, then segmentation accuracy is maintained, but the processing speed decreases

Engineering Contradiction:

Improvesegmentation accuracyVSAvoidprocessing speed

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent applies preliminary action by pre-calculating and storing distance values in a lookup table before the actual clustering process. The distance between pixels and representative data points is computed in advance and stored, so that during segmentation, only table lookup operations are needed instead of performing complex distance calculations, thereby reducing circuit complexity while maintaining accuracy

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent creates a copy of the distance information in the form of a lookup table that stores pre-computed distance values. Instead of recalculating distances during segmentation, the system copies and reuses these pre-stored values, dramatically speeding up the processing while ensuring the same accuracy as direct calculation would provide

Inventive Principle:

Principle #26Copying

3Productivity

If the search range for nearest representative data is limited to speed up processing, then processing speed increases, but measurement precision may be compromised

Engineering Contradiction:

Improveprocessing speedVSAvoiddistance calculation accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent segments the computational process into two distinct phases: an offline preprocessing phase where distance values are calculated and stored in a lookup table, and an online execution phase where only simple table lookups and comparisons are performed. This segmentation allows the complex calculation work to be done when computational resources are abundant, while the actual real-time segmentation uses minimal computational resources

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS9275466B2Image processing method and image processing apparatus for segmenting image into regions

Publication Date: 2016.03.01 CANON KK

AI summary

An image processing method includes, calculating a partial distance between a pixel of interest in an image and each of reference pixels, sequentially calculating a total distance between the pixel of interest and each of the plurality of the reference pixels based on the partial distance, determining a shortest total distance among the total distances that have been already calculated, in the sequential calculation of the total distance, and categorizing the pixel of interest based on the reference pixel corresponding to the shortest total distance, wherein, if the partial distance between the pixel of interest and a specific one of the reference pixels to be calculated is equal to or greater than the shortest total distance in the sequential calculation of the total distance, the calculation of the total distance between the pixel of interest and the specific one of the reference pixels to be calculated is omitted.