Conference proceeding
ADJUSTABLE ADABOOST CLASSIFIER AND PYRAMID FEATURES FOR IMAGE-BASED CERVICAL CANCER DIAGNOSIS
2015 IEEE 12TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING (ISBI), v 2015-, pp 281-285
01 Apr 2015
Featured in Collection : UN Sustainable Development Goals @ Drexel
Abstract
Cervical cancer is the third most common type of cancer in women worldwide. Most death cases of cervical cancer occur in less developed areas of the world. In this work, we develop an automated and low-cost method that is applicable in those low-resource regions. First, we propose a more distinctive multi-feature descriptor for encoding the cervical image information by enhancing an existing descriptor with the pyramid histogram of local binary pattern (PLBP) feature. Second, we apply the AdaBoost algorithm to perform feature selection, and train a binary classifier to differentiate high-risk patient visits from low-risk patient visits. Our AdaBoost classifier can be adjusted to achieve high specificity, which is necessary for use in clinical practice. Experiments on both balanced and imbalanced datasets are conducted to evaluate the effectiveness of our method. Our method is shown to achieve better performance than existing image-based CIN classification systems and also outperform human interpretations on various screening tests.
Metrics
Details
- Title
- ADJUSTABLE ADABOOST CLASSIFIER AND PYRAMID FEATURES FOR IMAGE-BASED CERVICAL CANCER DIAGNOSIS
- Creators
- Tao Xu - Lehigh UniversityEdward Kim - Villanova UniversityXiaolei Huang - Lehigh University
- Publication Details
- 2015 IEEE 12TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING (ISBI), v 2015-, pp 281-285
- Series
- IEEE International Symposium on Biomedical Imaging
- Publisher
- IEEE
- Number of pages
- 5
- Resource Type
- Conference proceeding
- Language
- English
- Academic Unit
- Computer Science (Computing)
- Web of Science ID
- WOS:000380546000067
- Scopus ID
- 2-s2.0-84944314759
- Other Identifier
- 991021884692704721
UN Sustainable Development Goals (SDGs)
This publication has contributed to the advancement of the following goals:
InCites Highlights
Data related to this publication, from InCites Benchmarking & Analytics tool:
- Collaboration types
- Domestic collaboration
- Web of Science research areas
- Engineering, Biomedical
- Engineering, Electrical & Electronic
- Radiology, Nuclear Medicine & Medical Imaging