Adaptive Compression and Reconstruction for Multidimensional Medical Image Data: A Hybrid Algorithm for Enhanced...

Area of Science:

• Medical Imaging
• Image Processing
• Data Compression

Background:

• Medical images (MI) require high fidelity during compression due to critical diagnostic details.
• Standard compression methods may degrade essential information in regions of interest (ROI).

Purpose of the Study:

• To develop an adaptive compression algorithm for medical images that preserves diagnostic quality.
• To effectively compress multi-dimensional medical image data from MRI, CT, and X-ray.

Main Methods:

• Image enhancement using Edge Enhancement-Contrast Limited Adaptive Histogram Equalization (EE-CLAHE) and adaptive anisotropic diffusion.
• Segmentation of images into ROI and non-ROI using Adaptive Expectation Maximization Clustering (AEMC), optimized with Fuzzy C-Means (FCM) and Otsu thresholding.
• Application of distinct compression schemes (Coiflet+Haar, Coiflet+Daubechies, modified SPIHT Huffman, EZW, SPIHT) to ROI and non-ROI.

Main Results:

• The combination of modified SPIHT Huffman for ROI and EZW for non-ROI demonstrated superior reconstruction quality.
• The proposed algorithm effectively preserves diagnostic details while reducing medical image dimensions.
• Experimental validation across MRI, CT, and X-ray modalities confirmed the algorithm's efficacy.

Conclusions:

• Adaptive compression is crucial for maintaining diagnostic integrity in medical imaging.
• The developed algorithm offers an effective solution for compressing diverse medical image modalities.
• Optimized compression strategies for ROI and non-ROI enhance overall image data management and analysis.