Related Experiment Video
Updated: Dec 15, 2025
11:01
Automated Imaging and Analysis for the Quantification of Fluorescently Labeled Macropinosomes
Published on: August 24, 2021
3.1K
Automated sperm morphology analysis approach using a directional masking technique
Hamza Osman Ilhan1, Gorkem Serbes2, Nizamettin Aydin1
1Department of Computer Engineering, Yildiz Technical University, Turkey.
Computers in Biology and Medicine
|July 14, 2020
Summary
Computer-based sperm morphology analysis improves accuracy over manual methods. A new framework using advanced preprocessing and machine learning enhances sperm image classification, boosting accuracy and efficiency.
Area of Science:
- Computer Vision
- Biomedical Image Analysis
- Machine Learning
Background:
- Sperm morphology assessment is crucial for evaluating sperm quality.
- Manual assessment is prone to human error and subjectivity.
- Computer-based automated analysis offers objective and reproducible results.
Purpose of the Study:
- To propose a novel computational framework for automated sperm morphology classification.
- To enhance the accuracy and efficiency of sperm image analysis.
- To overcome limitations of manual sperm assessment and existing automated methods.
Main Methods:
- A multi-stage cascade preprocessing pipeline including wavelet-based de-noising, shrinkage, gradient, and masking.
- Region-based descriptor features for image characterization.
- Non-linear kernel Support Vector Machine (SVM) for classification.
- Evaluation on the Human Sperm Head Morphology (HuSHeM) and Sperm Morphology Image Data Set (SMIDS).
Main Results:
- Cascaded preprocessing techniques significantly improved classification accuracy.
- Accuracy increased by 10% on HuSHeM and 5% on SMIDS datasets.
- The proposed framework outperformed most state-of-the-art methods.
- Achieved higher accuracy with reduced manual intervention (orientation, cropping).
Conclusions:
- The developed computational framework provides an accurate and efficient method for sperm morphology assessment.
- Automated analysis using advanced image processing and machine learning is superior to manual methods.
- The framework offers practical advantages by eliminating time-consuming manual operations.
