Related Experiment Video
Updated: Aug 4, 2025
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
42.7K
Cervical Transformation Zone Segmentation and Classification based on Improved Inception-ResNet-V2 Using Colposcopy
Srikanta Dash1, Prabira Kumar Sethy1, Santi Kumari Behera2
1Department of Electronics, Sambalpur University, Sambalpur, Odisha, India.
Cancer Informatics
|April 3, 2023
Summary
This study introduces a two-phase method using improved Inception-ResNet-v2 and SVM to segment and classify cervical transformation zones for accurate cervical cancer detection.
Area of Science:
- Oncology
- Medical Imaging
- Computer Vision
Background:
- Cervical cancer is the second most common malignancy in women globally.
- The cervical transformation zone is the primary site for aberrant cell development.
- Accurate identification of cervical cancer subtypes is crucial for effective treatment.
Purpose of the Study:
- To propose a novel two-phase method for segmenting and classifying the cervical transformation zone.
- To enhance the accuracy of cervical cancer identification using deep learning techniques.
- To improve the diagnostic capabilities for cervical cancer detection.
Main Methods:
- Image segmentation of the cervical transformation zone from colposcopy images.
- Application of data augmentation and an improved Inception-ResNet-v2 model for feature extraction.
- Multi-scale feature fusion using 3x3 convolution kernels from Reduction-A and Reduction-B, followed by Support Vector Machine (SVM) classification.
Main Results:
- The proposed model achieved 81.24% accuracy and 81.24% sensitivity.
- Specificity reached 90.62%, with a precision of 87.52%.
- The F1 score was 81.68%, indicating robust performance in classifying cervical cancer subtypes.
Conclusions:
- The developed two-phase method effectively segments and classifies the cervical transformation zone.
- The integration of Inception-ResNet-v2 with multi-scale feature fusion and SVM classification shows significant potential for improving cervical cancer diagnosis.
- This approach offers a promising tool for early and accurate detection of cervical cancer.
