Related Experiment Videos
Noise content analysis in clinical digital images
K S Chuang1, B J Liu, H K Huang
1Institute of Nuclear Science, National Tsing-Hua University, Taiwan.
Summary
Researchers used the Moran test to analyze noise in medical images. Removing noise bits from computed tomographic, magnetic resonance, and digital radiographic images did not affect overall image quality.
Area of Science:
- Medical Imaging
- Digital Signal Processing
- Radiology
Background:
- Digital radiological images typically use 12-bit precision.
- The presence of noise within these bits can impact diagnostic accuracy.
- Assessing the information content of each bit is crucial for image optimization.
Purpose of the Study:
- To evaluate the noise level in digital radiological images using a statistical method.
- To determine if all 12 bits of precision contain useful diagnostic information.
- To demonstrate that noise reduction techniques do not degrade image quality.
Main Methods:
- The Moran test was employed to quantify noise levels in computed tomographic (CT), magnetic resonance (MR), and digital radiography (DR) images.
- The test was applied to individual bit planes of each pixel.
- Image enhancement was performed on images after removing identified noise bits.
Main Results:
- Noise levels were successfully estimated for CT, MR, and DR images.
- Pixel data were effectively separated into signal and noise components.
- Preliminary results indicated no discernible difference in image quality between original and noise-bits-removed images.
Conclusions:
- A significant portion of the 12 bits in digital radiological images may represent noise rather than signal.
- Noise bit removal is a viable technique for image optimization without compromising diagnostic quality.
- This finding has potential implications for efficient storage and processing of medical imaging data.