… non-negative matrix deconvolution (NMD) which models spectrogram matrix as a linear
combination of dictionary … While the speaker recognition accuracy is lower for the new approach, …

… In this sense, with the aim of validating its use as a new limited-data automatic speaker
verification technique, this paper proposes to use the BLPOC function as a simple matching …

… Some of these methods make use of non-negative … spectrogram. In order to evaluate whether
a given parameter β1 is good for dictionary building, we take a reverberant spectrogram Y , …

… As a solution to the limited training data problem, nonnegative … (2) decomposes an source
spectrogram into A × ˆH where ˆH … all speaker’s voice into PPGs, and construct dictionaries that …

… Mel spectrogram representation of the elite dictionary of sub-… Wang et al. consider both
normalized cut and non-negative … features for five-way speaker identification. In particular, they …

… joint dictionary … spectrogram that prepares three parts like real, imaginary and magnitude
for each subband signal. Next, we utilize the GJDL approach for making the joint dictionaries, …

… 2D deconvolution layers to estimate the speech and noise … enhancement and noise-robust
speaker verification,” in Proc. … Venkataramani, “A neural network alternative to nonnegative …

… To obtain the aforementioned reconstructed voice spectrogram E 0 from F0, we define …
showed better results than supervised methods which use online dictionary learning (LRR, LRRi, …

… first solution to solve the speaker verification problem when the test … the non-negative
spectrogram Y ∈ RF×T , as follows, … a basis dictionary by decomposing the clean spectrogram Sc …

… Smaragdis [27] introduced NMFD (non-negative matrix factor deconvolution), which is
NMF … of instruments inside the spectrogram, we need an algorithm to approximate a non-negative …