Original Text:
We have demonstrated signal feature extraction technique of citotoxicity testing using wavelet packet decompositions. The non stationary signals from impedemetric measurement were extracted and analysed at level 3 and we found that each signal has their own signi cant characteristics. In addition we applied statistical features to gather their features for unsupervised learning algorithm input vectors. The technique revealed that in certain subband, each signal has their own statistical features depending on concentration of H2O2 in the well.
We also have validated the results with photography method and provide connected results. Further study is required with di erent type of toxin to determine possibly signi cant output
as comparison with our result in this paper.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Revised Text:
We have demonstrated a signal feature extraction technique of cito-toxicity testing, using wavelet packet decompositions. The non stationary signals from impedemetric measurement were extracted and analysed at level 3, and we found that each signal has its own signifi cant characteristics. In addition we applied statistical methods to establish their features for unsupervised learning algorithm input vectors. This technique revealed, that in certain sub-bands, each signal has its own statistical features, depending on concentration of H2O2 in the well.
We have validated the results with the photography method and provide connected results. Further study is required with diff erent types of toxins to determine possible significant outputs in comparison with our results in this paper.
Editor:
1. Please review the text carefully to ensure your intended meaning was not mistakenly changed.
2. Before you leave this page, please take a moment to use the "Click here to post comments." link at the bottom of the page to rate the proofreading and leave a comment if you wish.
Proofreading and editing pages that receive ratings and comments are moved to the top of the list.