Entropy improvement by the Temporal-Window method for alternating and non-alternating 3D wavelet transform over...

Area of Science:

• Medical Imaging
• Signal Processing
• Data Compression

Background:

• Three-dimensional wavelet transform (3D-WT) offers lossless coding and high-quality reconstruction, crucial for medical imaging.
• Existing 3D-WT methods have limitations in efficiently processing volumetric data, particularly temporal redundancies.

Purpose of the Study:

• To introduce and evaluate four novel algorithms based on the Classic 3D-WT for volumetric data coding.
• To enhance the efficiency of 3D-WT by improving spatial and temporal decomposition processes.
• To reduce data redundancies and achieve lower entropy in medical imaging sequences.

Main Methods:

• Development of four new algorithms building upon the Classic 3D-WT.
• Implementation of the Temporal-Window method for temporal decomposition.
• Experimental evaluation using a representative dataset of angiography sequences.

Main Results:

• The proposed algorithms successfully obtain wavelet coefficients after spatial and temporal decomposition.
• Significant reduction in redundancies within the video sequences was observed.
• Lower entropy values were achieved compared to the Classic 3D-WT algorithm.

Conclusions:

• The novel 3D-WT algorithms demonstrate improved performance in volumetric data coding for medical applications.
• The Temporal-Window method effectively reduces redundancies and enhances coding efficiency.
• These findings are particularly relevant for processing intrinsically volumetric medical imaging modalities like angiography.