Impact of Image Dynamic Range on Brain Texture Classification
Author Information
Author(s): Mahmoud-Ghoneim Doaa, Alkaabi Mariam K, de Certaines Jacques D, Goettsche Frank-M
Primary Institution: United Arab Emirates University
Hypothesis
How does the dynamic range of image greylevels affect the classification of brain white matter using the Cooccurrence Matrix method?
Conclusion
The dynamic range used for Cooccurrence Matrix calculation significantly influences the classification results for brain white matter.
Supporting Evidence
- The best classification results were achieved with a dynamic range of 128 greylevels.
- Larger dynamic ranges can increase computation costs and limit method performance.
- The study found that classification accuracy is highly dependent on the dynamic range of image quantization.
Takeaway
Changing the number of colors in an image can help computers tell different parts of the brain apart better, but too many colors can make it harder.
Methodology
MR images were obtained from glioblastoma patients, and two calculation approaches for the Cooccurrence Matrix were used on different dynamic ranges.
Limitations
The study only included glioblastoma patients and may not generalize to other conditions.
Participant Demographics
Patients with histologically confirmed brain glioblastoma, average age 53 ± 18.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website