### Table 1: Blind source separation results of speech mixtures (a) Noiseless

"... In PAGE 3: ... However, the Euler-like integration preserves the orthogonalityin a satisfactory way as well. Table1 lists results of blind source separation of speech mix- tures. Twenty speech signals, each of 3,500 samples, were mixed by a randomly generated mixing matrix.... ..."

### Table 2. BKYY DR for ICA and Blind Source Separation (BSS)

1997

Cited by 19

### Table 1: Summary of procedures for blind source separation using generalized eigenvalue decomposition. Given a N T matrix X containing

2003

"... In PAGE 3: ...diagonal cross-statistics Q. A summary of the different assumptions and choices for Q is given in Table1 . We also show experimental results for signals that simultaneously satisfy the various assumptions.... In PAGE 7: ...g. see Table1 ), we believe that it is a simple framework for understanding and comparing the various approaches, as well as a method for verifying the underlying statistical assumptions. Acknowledgments We would like to thank Clay Spence and Craig Fancourt for their help and discussions.... ..."

Cited by 7

### Table 1: The convolutive mixing equation and its corresponding separation equation are shown for different domains in which blind source separation algorithms have been derived. Mixing Process Separation Model

"... In PAGE 4: ... (15) As with the mixing process, the separation system can be expressed in the z-domain as Y (z) = W (z)X(z), (16) and it can also be expressed in block Toeplitz form with the corresponding definitions for hatwide y(t) and hatwider W [25]: hatwide y(t) = hatwider W hatwide x(t). (17) Table1 summarizes the mixing and separation equations in the different domains. 3.... ..."

### Table 1: Mixing model and ICA solution

2006

"... In PAGE 1: ... 2) The use of all the information obtained from the basis vectors solves the permutation problem more accurately and therefore improves the BSS performance. Blind source separation in frequency domain Table1 shows equations related to a mixing model and ICA. Convolutive mixtures in the time domain can be approximated as multiple instantaneous mixtures in the frequency domain.... ..."

Cited by 2

### Table 3: Probability of retrieval of artefact classes. The value one means that there is at least one component that represents this class for every run of the algorithm.

"... In PAGE 19: ...Table3 , the results are summarised for each algorithm and artefact class. The abbreviations of the classes are explained in the caption of Figure 8; avg.... In PAGE 19: ... It can be noted that none of the ICA algorithms is best for all the classes. [ Table3 about here.] 7 Conclusions In this paper, we have experimentally compared ve prominent neural or semi-neural learning algorithms introduced for independent component analysis or blind source separation.... ..."

### TABLE III COMPARISON OF THE BLIND COOPERATION SCHEME AND THE SEPARATION BASED SCHEME (THEORETICAL LIMITS) Separation Blind Cooperation

2004

Cited by 3

### Table 1: Distribution of Blind Test Material Among Sources

"... In PAGE 5: ...aterial, approx. 50k words, as blind test and used the remainder as training. For Hindi, a test set of 25 previously unseen documents, consisting of 50k total words, was common to all participating sites. Not only were the documents previously unseen, but their distribution among sources was drastically different than in training, as is evident in Table1 below. Therefore, though the genre (news) was the same as in ... ..."