3D-DCT Coding for Video Editing, Low-Power Devices, and Medical Imaging

This paper evaluates the practical use of the Three-Dimensional Discrete Cosine Transform (3D-DCT) for video and volumetric image compression. While one- and two-dimensional DCT transforms are widely used in modern multimedia standards, their three-dimensional extension has received limited attention in real coding systems. In this work, a complete 3D-DCT–based encoder is developed by extending a JPEG-like pipeline to operate on three-dimensional data blocks. The proposed approach processes groups of video frames as 3D cubes and applies a separable 3D-DCT followed by quantization, coefficient serialization, and entropy coding. Unlike conventional video codecs that rely on motion estimation and compensation, the proposed system exploits temporal redundancy directly through the transform domain, resulting in a simpler coding structure with reduced algorithmic complexity. Different configurations are evaluated, including alternative methods for computing the 3D quantization matrix, serialization schemes, and variable block depth along the temporal dimension.Experimental results obtained from multiple test videos and volumetric medical datasets, including CT and MRI studies, demonstrate that the proposed method achieves competitive compression ratios while maintaining good reconstruction quality measured in terms of Peak Signal-to-Noise Ratio (PSNR). The results indicate that 3D-DCT provides a flexible and computationally simple solution for both video compression and three-dimensional medical image coding, particularly in applications where implementation simplicity and frame-level accessibility are important.