Related Experiment Videos
Multiple resolution Bayesian segmentation of ultrasound images
1Department of Electrical Engineering, University of Rochester, NY 14627, USA.
Ultrasonic Imaging
|October 1, 1995
Summary
We developed a new method for segmenting ultrasound images with speckle noise. This technique improves accuracy and detail, especially for low-contrast areas, by converting Rayleigh to Gaussian statistics for faster processing.
Area of Science:
- Medical Imaging
- Image Processing
- Ultrasound Technology
Background:
- Speckle noise in ultrasound images degrades image quality and segmentation accuracy.
- Accurate segmentation is crucial for quantitative analysis and diagnosis in medical ultrasound.
Purpose of the Study:
- To propose a novel method for maximum a posteriori (MAP) probabilistic segmentation of speckle-laden ultrasound images.
- To improve the accuracy and low-contrast detail in ultrasound image segmentation.
Main Methods:
- A multiple-resolution based technique is employed.
- Speckle images with Rayleigh statistics are converted to subsampled images with Gaussian statistics.
- This conversion facilitates accurate parameter estimation for the segmentation algorithm.
Main Results:
- The proposed method demonstrates improved performance over existing techniques.
- Enhanced accuracy in segmenting low-contrast details within ultrasound images.
- Reduced computation time due to the statistical conversion process.
Conclusions:
- The novel method offers a significant advancement in ultrasound image segmentation.
- The technique effectively addresses challenges posed by speckle noise.
- It provides a more accurate and efficient approach for medical image analysis.