Automatic classification and detection of clinically-relevant images for diabetic retinopathy

Xinyu Xu, Baoxin Li

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

We proposed a novel approach to automatic classification of Diabetic Retinopathy (DR) images and retrieval of clinically-relevant DR images from a database. Given a query image, our approach first classifies the image into one of the three categories: microaneurysm (MA), neovascularization (NV) and normal, and then it retrieves DR images that are clinically-relevant to the query image from an archival image database. In the classification stage, the query DR images are classified by the Multi-class Multiple-Instance Learning (McMIL) approach, where images are viewed as bags, each of which contains a number of instances corresponding to non-overlapping blocks, and each block is characterized by low-level features including color, texture, histogram of edge directions, and shape. McMIL first learns a collection of instance prototypes for each class that maximizes the Diverse Density function using Expectation-Maximization algorithm. A nonlinear mapping is then defined using the instance prototypes and maps every bag to a point in a new multi-class bag feature space. Finally a multi-class Support Vector Machine is trained in the multi-class bag feature space. In the retrieval stage, we retrieve images from the archival database who bear the same label with the query image, and who are the top K nearest neighbors of the query image in terms of similarity in the multi-class bag feature space. The classification approach achieves high classification accuracy, and the retrieval of clinically-relevant images not only facilitates utilization of the vast amount of hidden diagnostic knowledge in the database, but also improves the efficiency and accuracy of DR lesion diagnosis and assessment.

Original languageEnglish (US)
Title of host publicationProgress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume6915
DOIs
StatePublished - 2008
EventMedical Imaging 2008 - Computer-Aided Diagnosis - San Diego, CA, United States
Duration: Feb 19 2008Feb 21 2008

Other

OtherMedical Imaging 2008 - Computer-Aided Diagnosis
CountryUnited States
CitySan Diego, CA
Period2/19/082/21/08

Fingerprint

Probability density function
Support vector machines
Labels
Textures
Color

Keywords

  • Computer-aided diagnosis
  • Diabetic retinopathy lesion detection and classification
  • Image retrieval
  • Multiple-instance learning

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Xu, X., & Li, B. (2008). Automatic classification and detection of clinically-relevant images for diabetic retinopathy. In Progress in Biomedical Optics and Imaging - Proceedings of SPIE (Vol. 6915). [69150Q] https://doi.org/10.1117/12.769858

Automatic classification and detection of clinically-relevant images for diabetic retinopathy. / Xu, Xinyu; Li, Baoxin.

Progress in Biomedical Optics and Imaging - Proceedings of SPIE. Vol. 6915 2008. 69150Q.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Xu, X & Li, B 2008, Automatic classification and detection of clinically-relevant images for diabetic retinopathy. in Progress in Biomedical Optics and Imaging - Proceedings of SPIE. vol. 6915, 69150Q, Medical Imaging 2008 - Computer-Aided Diagnosis, San Diego, CA, United States, 2/19/08. https://doi.org/10.1117/12.769858
Xu X, Li B. Automatic classification and detection of clinically-relevant images for diabetic retinopathy. In Progress in Biomedical Optics and Imaging - Proceedings of SPIE. Vol. 6915. 2008. 69150Q https://doi.org/10.1117/12.769858
Xu, Xinyu ; Li, Baoxin. / Automatic classification and detection of clinically-relevant images for diabetic retinopathy. Progress in Biomedical Optics and Imaging - Proceedings of SPIE. Vol. 6915 2008.
@inproceedings{21a045ecdad645e08618e4a08b1028a7,
title = "Automatic classification and detection of clinically-relevant images for diabetic retinopathy",
abstract = "We proposed a novel approach to automatic classification of Diabetic Retinopathy (DR) images and retrieval of clinically-relevant DR images from a database. Given a query image, our approach first classifies the image into one of the three categories: microaneurysm (MA), neovascularization (NV) and normal, and then it retrieves DR images that are clinically-relevant to the query image from an archival image database. In the classification stage, the query DR images are classified by the Multi-class Multiple-Instance Learning (McMIL) approach, where images are viewed as bags, each of which contains a number of instances corresponding to non-overlapping blocks, and each block is characterized by low-level features including color, texture, histogram of edge directions, and shape. McMIL first learns a collection of instance prototypes for each class that maximizes the Diverse Density function using Expectation-Maximization algorithm. A nonlinear mapping is then defined using the instance prototypes and maps every bag to a point in a new multi-class bag feature space. Finally a multi-class Support Vector Machine is trained in the multi-class bag feature space. In the retrieval stage, we retrieve images from the archival database who bear the same label with the query image, and who are the top K nearest neighbors of the query image in terms of similarity in the multi-class bag feature space. The classification approach achieves high classification accuracy, and the retrieval of clinically-relevant images not only facilitates utilization of the vast amount of hidden diagnostic knowledge in the database, but also improves the efficiency and accuracy of DR lesion diagnosis and assessment.",
keywords = "Computer-aided diagnosis, Diabetic retinopathy lesion detection and classification, Image retrieval, Multiple-instance learning",
author = "Xinyu Xu and Baoxin Li",
year = "2008",
doi = "10.1117/12.769858",
language = "English (US)",
isbn = "9780819470997",
volume = "6915",
booktitle = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",

}

TY - GEN

T1 - Automatic classification and detection of clinically-relevant images for diabetic retinopathy

AU - Xu, Xinyu

AU - Li, Baoxin

PY - 2008

Y1 - 2008

N2 - We proposed a novel approach to automatic classification of Diabetic Retinopathy (DR) images and retrieval of clinically-relevant DR images from a database. Given a query image, our approach first classifies the image into one of the three categories: microaneurysm (MA), neovascularization (NV) and normal, and then it retrieves DR images that are clinically-relevant to the query image from an archival image database. In the classification stage, the query DR images are classified by the Multi-class Multiple-Instance Learning (McMIL) approach, where images are viewed as bags, each of which contains a number of instances corresponding to non-overlapping blocks, and each block is characterized by low-level features including color, texture, histogram of edge directions, and shape. McMIL first learns a collection of instance prototypes for each class that maximizes the Diverse Density function using Expectation-Maximization algorithm. A nonlinear mapping is then defined using the instance prototypes and maps every bag to a point in a new multi-class bag feature space. Finally a multi-class Support Vector Machine is trained in the multi-class bag feature space. In the retrieval stage, we retrieve images from the archival database who bear the same label with the query image, and who are the top K nearest neighbors of the query image in terms of similarity in the multi-class bag feature space. The classification approach achieves high classification accuracy, and the retrieval of clinically-relevant images not only facilitates utilization of the vast amount of hidden diagnostic knowledge in the database, but also improves the efficiency and accuracy of DR lesion diagnosis and assessment.

AB - We proposed a novel approach to automatic classification of Diabetic Retinopathy (DR) images and retrieval of clinically-relevant DR images from a database. Given a query image, our approach first classifies the image into one of the three categories: microaneurysm (MA), neovascularization (NV) and normal, and then it retrieves DR images that are clinically-relevant to the query image from an archival image database. In the classification stage, the query DR images are classified by the Multi-class Multiple-Instance Learning (McMIL) approach, where images are viewed as bags, each of which contains a number of instances corresponding to non-overlapping blocks, and each block is characterized by low-level features including color, texture, histogram of edge directions, and shape. McMIL first learns a collection of instance prototypes for each class that maximizes the Diverse Density function using Expectation-Maximization algorithm. A nonlinear mapping is then defined using the instance prototypes and maps every bag to a point in a new multi-class bag feature space. Finally a multi-class Support Vector Machine is trained in the multi-class bag feature space. In the retrieval stage, we retrieve images from the archival database who bear the same label with the query image, and who are the top K nearest neighbors of the query image in terms of similarity in the multi-class bag feature space. The classification approach achieves high classification accuracy, and the retrieval of clinically-relevant images not only facilitates utilization of the vast amount of hidden diagnostic knowledge in the database, but also improves the efficiency and accuracy of DR lesion diagnosis and assessment.

KW - Computer-aided diagnosis

KW - Diabetic retinopathy lesion detection and classification

KW - Image retrieval

KW - Multiple-instance learning

UR - http://www.scopus.com/inward/record.url?scp=44349150338&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=44349150338&partnerID=8YFLogxK

U2 - 10.1117/12.769858

DO - 10.1117/12.769858

M3 - Conference contribution

AN - SCOPUS:44349150338

SN - 9780819470997

VL - 6915

BT - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

ER -