Separable Complex-Valued Graph Filter Banks for Graph Signals: A general method for the design of orthogonal and biorthogonal separable complex-valued filter banks on graphs is proposed, in which the complex filters are obtained directly from the existing real filters.
Chunxu GuoYulong QiaoPublished in: ICDSP (2021)
Keyphrases
- complex valued
- filter bank
- infinite impulse response
- wavelet filter banks
- lifting scheme
- perfect reconstruction
- fir filters
- lapped transforms
- filter coefficients
- finite impulse response
- real valued
- image coding
- synthesis filters
- subband
- digital filters
- multiresolution
- wavelet filters
- signal processing
- vanishing moments
- recurrent neural networks
- wavelet transform
- multiscale
- filter design
- image compression
- computationally efficient
- reconstruction error
- compactly supported
- wavelet packet
- low pass
- linear equations
- multiresolution analysis
- recursive least squares
- wavelet decomposition
- discrete wavelet transform
- filter responses
- frequency domain
- activation function
- discrete cosine transform
- wavelet bases
- quadrature mirror
- graph structure
- blind source separation
- wavelet basis
- low pass filter
- support vector machine
- computer vision
- feature vectors
- image processing