Blob Detection Using Closed Curve Intensity Analysis

Solution Overview

Problem

Existing blob detection techniques are inadequate for accurately detecting blobs with specific shapes under varying lighting conditions and are computationally expensive, leading to low accuracy and high false positives in applications like optical motion capture and photogrammetry.

Innovation Solution

A method that derives detection scores by analyzing the proportion of rays crossing a closed curve with a predetermined shape, where the intensity differential exceeds a contrast threshold, allowing for rapid and accurate detection of blobs by focusing on specific shapes rather than general regions of high contrast.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If template matching and matched filter techniques are used to detect blobs, then detection accuracy can be improved for specific shapes, but computation time increases significantly

Engineering Contradiction:

Improveblob detection accuracyVSAvoidcomputation time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The detection process is segmented into two distinct stages: a fast elimination stage that quickly filters out non-blob regions using simple intensity thresholding, and a refined detection stage that applies shape-specific analysis only to candidate regions. This segmentation allows the system to maintain high accuracy for specific shapes while dramatically reducing overall computation time by avoiding exhaustive template matching across the entire image.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Instead of applying comprehensive template matching to all pixels, the invention applies partial action by first performing a rapid elimination of non-candidate regions using simple intensity-based criteria. Only regions that pass this preliminary filter undergo the more computationally intensive shape-specific detection, thus achieving accurate detection of specific shapes with reduced computation time.

Inventive Principle:

Principle #16Partial or excessive action

2Productivity

If simple thresholding is used for blob detection, then computation speed is high and real-time detection is achieved, but detection accuracy deteriorates under variant lighting conditions

Engineering Contradiction:

Improvedetection speedVSAvoidblob detection accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The detection process is segmented into two distinct stages: a fast elimination stage that quickly filters out non-blob regions using simple intensity thresholding, and a refined detection stage that applies shape-specific analysis only to candidate regions. This segmentation allows the system to maintain high accuracy for specific shapes while dramatically reducing overall computation time by avoiding exhaustive template matching across the entire image.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The invention applies different detection strategies to different regions of the image based on their likelihood of containing blobs. High-contrast regions undergo simple thresholding for speed, while candidate regions undergo more rigorous shape-specific analysis. This local differentiation of detection quality allows the system to optimize both speed and accuracy in appropriate contexts.

Inventive Principle:

Principle #3Local quality

3Reliability

If comprehensive templates are used to compensate for shape differences and noise, then detection reliability improves, but device complexity and computation time increase

Engineering Contradiction:

Improvedetection reliabilityVSAvoidtemplate set complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The detection process is segmented into two distinct stages: a fast elimination stage that quickly filters out non-blob regions using simple intensity thresholding, and a refined detection stage that applies shape-specific analysis only to candidate regions. This segmentation allows the system to maintain high accuracy for specific shapes while dramatically reducing overall computation time by avoiding exhaustive template matching across the entire image.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Instead of applying comprehensive template matching to all pixels, the invention applies partial action by first performing a rapid elimination of non-candidate regions using simple intensity-based criteria. Only regions that pass this preliminary filter undergo the more computationally intensive shape-specific detection, thus achieving accurate detection of specific shapes with reduced computation time.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS8103055B2Detection of blobs in images

Publication Date: 2012.01.24 OXFORD METRICS PLC

AI summary

Blobs are detected in an image using closed curves having a predetermined shape such as a circle. Positions in the image are analysed. The size is determined at which there is a maximum in the differential of the average intensity around a closed curve with respect to the size of the closed curve. Detection scores are derived, representing the proportion of rays, out of a plurality of rays crossing a closed curve of the determined size, along which the intensity differential across the closed curve exceeds a contrast threshold. Detection of blobs is performed on the basis of the detection scores exceeding a threshold. Pixels of a detected blob are segmented for calculation of a centroid using a blob-separation threshold which is the average intensity around the closed curve of the determined size. The technique allows accurate and rapid detection of blobs of the predetermined shape.