### Table 1: SNR, CNR and resolution measures for a noisy image, an average of two acquisitions, a

1996

"... In PAGE 9: ...Table 1: SNR, CNR and resolution measures for a noisy image, an average of two acquisitions, a 4-connected median #0Cltered image, a correlated image, a Wiener-correlated image and a 2x2 spatial averaged image. In Table1 the gain in SNR and CNR along with the estimated value for the loss of resolution is shown for a noisy image, the average of two acquisitions and for images processed with a 4 connected median #0Clter and the proposed correlation techniques. For comparison we included the results for a 2#022 spatial averaging #0Clter.... ..."

Cited by 1

### Table 7. Coefficients of Determination (Square Of Correlation Coefficients) for ESL and Measured Level for Various Metrics; Noisiness Judgements

"... In PAGE 19: ...6 Correlation of equivalent subjective level with objective measures. Table7 presents the results of linear regression analyses, based on the noisiness judgements, of the equivalent subjective level on measured sound level for the different metrics and binaural summation methods. Some differences in results of these analyses, as compared to the results in table 5 for D ESL, are noted.... ..."

### Table 7: Comparison of the proposed system with the noisy system using the log-likelihood ratio measure for the Blackhawk helicopter noise. Phone Class Noisy Baseline Proposed System

2005

### Table 5: Comparison of the proposed system with the noisy system using the log-likelihood ratio measure for the M2 tank noise. Phone Class Noisy Baseline Proposed System

2005

### Table 1: Comparison of different localization algorithms in RiSt; the averaged localization error is the distance between exact positions and the estimated positions which used measured noisy distances as input

### Table 1: An ideal population of edge pixels has been perturbed with Gaussian noise of di erent variance, as indicated by 2 above the three rightmost columns. The segment lengths and orientations have been computed for each of 1000 noisy version of the original population. The percentage of orientation/length measurement falling inside the ellipse of uncertainty (as determined by the presented noise model) is in the table compared to the percentage predicted by the 2 table, (leftmost column). This has been done for varying k2 parameters, (larger k2 means larger percentage of noisy measurements falling inside ellipse).

1995

Cited by 4

### Table 4: Comparison of the proposed system with the noisy system using the segmental SNR measure for the M2 tank noise. Highest SNR cases are shown with bold fonts. Phone Class Noisy (dB) Baseline (dB) Proposed System (dB)

2005

### Table 2. The recognition rates of noisy speech recorded in real environments (%)

1999

"... In PAGE 6: ... One of the loudspeakers played speech signals while the other played the sound of F-16 fighter noise signal. Table2 shows the recognition rates of noisy speech recorded in real environments. The higher measured SNR of two input channels was represented in the first column, and a test word and other ar- bitrary chosen 9 words were used to train the unmixing matrices.... In PAGE 9: ...Table 1. The recognition rates with white Gaussian noise according to the numbers of small bands (%) Table2 . The recognition rates of noisy speech recorded in real environments (%)... ..."

Cited by 7

### Table 2 Summary offunction approximation results using FasArt and FasBack with noisy or non-noisy learning patterns

"... In PAGE 10: ... Furthermore, a widely accepted quantitative measure, such as RMSE (Relative Mean Square Error), has been used in order to draw valid conclusions. In the results, shown in Table2 , it can be seen that FasBack is better than FasArt in both cases for equivalent network complex- ity. It is noteworthy that difference in performance (predic- tion error) is not significant when noise is incorporated into the learning patterns.... ..."

### Table 5 shows the results. A comparison of the first two runs shows that using noisy data on training degrades the performance. In the third run, where all models were considered more uniformly, the performance improved consistently on all measures: precision, recall, F-measure, accuracy and AUC.

"... In PAGE 7: ... Table5 : IAS Results: precision (P), recall (R), accuracy (A), F-score (F), AUC 5 Protein Interaction Pairs (IPS) For the IPS sub-task, given a set of full text articles, we were asked to produce for each one a ranked list of interacting UniProt IDs. We built a classifier, which, given a pair of UniProt IDs, from the same organism and co-occurring in the same sentence, decides on whether they interact or not.... ..."