Abstract:The Karhunen-Loève transform (KLT) stands as a well-established discrete transform, demonstrating optimal characteristics in data decorrelation and dimensionality reduction. Its ability to condense energy compression into a select few main components has rendered it instrumental in various applications within image compression frameworks. However, computing the KLT depends on the covariance matrix of the input data, which makes it difficult to develop fast algorithms for its implementation. Approximations for the KLT, utilizing specific rounding functions, have been introduced to reduce its computational complexity. Therefore, our paper introduces a category of low-complexity, data-independent KLT approximations, employing a range of round-off functions. The design methodology of the approximate transform is defined for any block-length $N$, but emphasis is given to transforms of $N = 8$ due to its wide use in image and video compression. The proposed transforms perform well when compared to the exact KLT and approximations considering classical performance measures. For particular scenarios, our proposed transforms demonstrated superior performance when compared to KLT approximations documented in the literature. We also developed fast algorithms for the proposed transforms, further reducing the arithmetic cost associated with their implementation. Evaluation of field programmable gate array (FPGA) hardware implementation metrics was conducted. Practical applications in image encoding showed the relevance of the proposed transforms. In fact, we showed that one of the proposed transforms outperformed the exact KLT given certain compression ratios.

Fast Binary Block Inverse Jacket Transform

Fast Block Jacket Transform Based on Pauli Matrices

A Generalized Reverse Block Jacket Transform.

Block-Circulant Inverse Orthogonal Jacket Matrices and Its Applications to the Kronecker MIMO Channel

A High Performance Inverse Integer Transform Architecture for H. 264

Fast Reciprocal Jacket Transform With Many Parameters

Inverse Transform Architecture for VC1 Adaptive Block Size

Fast parametric reciprocal-orthogonal jacket transforms

The Inverses of Block Tridiagonal Matrices

The Inverses of Block Toeplitz Matrices

Orthogonal Space-Time Block Codes Design Using Jacket Transform for MIMO Transmission System

Computation of Outer Inverse of Tensors Based on t‐Product

Blockwise inversion and algorithms for inverting large partitioned matrices

Fast Data-independent KLT Approximations Based on Integer Functions

Error-resistance and Low-complexity Integer Inverse Discrete Cosine Transform

Fast Algorithm and Architecture Design for H.264/AVC Multiple Transforms

Algorithms for Finding the Inverses of Factor Block Circulant Matrices

A note on the group inverse of anti-triangular block matrices over skew filed

Randomized Block Kaczmarz Methods for Inner Inverses of a Matrix

Intra-predictive Transforms for Block-Based Image Coding.