GDV Measures Vitality?
Igor Kononenko, Miha Sedej, Aleksander Sadikov
University of Ljubljana, Faculty of Computer and Information Science
{igor.kononenko; aleksander.sadikov}@fri.uni-lj.si
Abstract
The study verifies a hypothesis that GDV in fact measures vitality. For that purpose a limited study was performed and the findings support the hypothesis. However, further investigation is needed in order to more reliably confirm it.
1. Motivation
The term “vitality” is usually referring to the ability or capacity to live, the ability to exist, the strength of a living organism, or even to the energy for living. Some authors use vitality as a synonym for health. Although health may not be appropriate synonym of vitality, they are definitely highly related. Therefore any device that could measure vitality would be measuring also at least some essential aspects of the health status of the living organism. Gas discharge visualization methodology (Korotkov, 2002) seems a good candidate for a general device that is able to measure vitality. Up to date, many studies have been completed and published that show various aspects of the ability of GDV methodology to measure certain aspects of health status and vitality (see Korotkov, 2004). The aim of the study, presented in this paper, is to further verify the hypothesis stated in the title of the paper. In order to test the ability to measure vitality, one has to select an appropriate controllable scenario. Recording of humans highly depends on many parameters which are hard to control and even if one would be able to control the parameters of the environment, the interpretation of results is still questionable, as the influence of human psyche cannot be controlled to the satisfactory extent. Therefore we decided to record plants which are much easier to control, although we are aware that even with plants many of the environmental parameters cannot be controlled in such a limited study. We already have some experience with recording plants with GDV (Sadikov et al., 2004), however in previous studies we did not try to measure vitality per se.
For this study we selected a well defined sub-problem of the influence of the UV-light on the vitality of plants. This problem is well controllable and the influence of UV-light can be at least to some extent reliably hypothesized. Therefore we can compare the expected behavior with the actual measures performed by GDV methodology. On the other hand, however, the study may reveal also unknown influence of UV-light on vitality. In such a case, a vast amount of biological knowledge can assist in the investigation and explanation of the phenomena. In the next section the experimental design and the procedure of recording the coronas of pine needles are described. Section 3 gives results of the recordings and of the statistical analysis and Section 4 concludes.
2. Experimental design
We selected to test the influence of UV-light on the vitality of pines. For that purpose we decided to record the coronas of needles of pines (Picea abies (L.) Karst.) that were exposed to normal amount and extra amount of UV-light. The study was further expanded to test the influence of UV-light in different temperatures. Figure 1 gives a summary of recordings. The group Control are pines under normal UV-light with an addition of extra UV-A light. The Plus group are pines with extra UV-A and UV-B light. Therefore, the difference between Control and Plus is in the amount of UV-B light. Pines under Warm conditions were recorded at the end of November 2003 (average temperature +10,7°C) and under Cold conditions at the end of January 2004 (average temperature -2,5°C). Other weather conditions were approximately the same for Warm and Cold recording. Pines were 6 years old and were already for 5 years exposed to extra UV-light. The needles of pines are of different age. We used three different ages of needles: C - Current year; C+1 - from last year (one year old); C+2 - two years old.
Figure 1: The number of recorded pine needles under different conditions.
Figure 2: Classification accuracy when classifying new needles of various age (C, C+1, C+2) and at different temperatures (Warm - left three bars, Cold - right three bars) into classes Control and Plus (prior probability of correct classification is 50%).
3. Results
We recorded the coronas of needles with GDV device, developed by prof. Korotkov and his team (Korotkov, 1998; 2002). We used GDV Assistant sowtware package for parametrization of digital pictures, developed by our team (Sadikov, 2002). The numerical parameters (features, attributes) were then analysed using machine learning techniques (see e.g. Kononenko, 2001) and statistical analysis. In particular, we used decision trees generated by C4.5 algorithm (Quinlan, 1993). Results are summarized in Figures 2, 3, and 4.
Figure 3: Classification accuracy when classifiying new needless in one of three age classes (C, C+1 and C+2) for Warm weather (left two bars) and Cold weather (right two bars); K = Control, P = Plus; prior probability of correct classification is 33.3%.
Figure 4: Classification accuracy when classifying new needles in classes Warm and Cold (prior probability of correct classification is 50%).
In Figure 2 we can see that the differences between Control and Plus groups of needles are easier to detect during Cold temperatures. It is obvious that extra UV-B light has strong influence on pines and Figure 2 shows that this difference can be detected with GDV device. The problem with Warm weather is in greater variance which makes classification more difficult. It is also obvious that the influence of UV-light has different influence on needles of different age. Figure 3 shows that GDV technology is able to reliably differentiate between needles of different age. Figure 2 and also indicates the opposite effect of extra UV-light on older needles when compared to young needles. This was further verified with statistical analysis (not described in this summary). Figure 4 shows that GDV is also able to differentiate well between Cold and Warm weather.
4. Conclusions
The results show, that the extra UV-B light is stressful for the plant and its effect is different on young and old needles. The visualization of the effect with GDV technology depends on the weather conditions, in particular it is different for warm and cold weather. In cold conditions the variability of measurements is much lower which is most likely caused by the lower activity of the plants during winter. Statistical analysis (not given in this abstract) has shown that older and younger needles show the opposite trend which can be explained with different effect of extra UV-light. For current (young) needles the extra UV-light has positive effect, however, for older needles it is harmful.
It is interesting that we obtained great differences between needles of different age. Older needless exhibit higher variance for one of the numerical attributes (form coefficient). We can conclude that our study using GDV device shows expected behaviour of pine needles with respect to different amount of extra UV-B light and with respect to different age of needles and with respect to different outside temperature. We were even able to relate some numerical parameters, extracted from digital images of coronas, with various aspects of vitality status of the pines. Therefore, our study supports the conclusion that GDV measures at least some aspects of vitality. This hypothesis, however, needs additional empirical evidence.
Acknowledgements
We thank prof. dr. Alenka Gaberscik for providing the experimental pines and in dr. Joze Bavcon for his help and support.
References
[1] I. Kononenko, Machine learning for medical diagnosis: History, state of the art and perspective, Artificial Intelligence in Medicine, vol. 23, no. 1, 2001, pp. 89-109.
[2] K. Korotkov, Aura and Consciousness, State Editing & Publishing Unit “Kultura”, St. Peterburg, 1998.
[3] K. Korotkov, Human Energy Field: Study with GDV Bioelectrography, Backbone Publishing, Fair Lawn, NJ, USA, 2002.
[4] K. Korotkov (ed.), Measuring Energy Fields: State of the Science, Backbone Publishing, Fair Lawn, NJ, USA, 2004.
[5] J.R. Quinlan, C4.5 Programs for Machine Learning, Morgan Kaufmann, 1993.
[6] A. Sadikov, Computer visualization, parametrization and analysis of GDV images, M.Sc. Thesis, University of Ljubljana, Faculty of Computer and Information Science, Ljubljana, Slovenia, 2002 (in Slovene).
[7] A. Sadikov, I. Kononenko, F.. Weibel, Analyzing Coronas of Fruits and Leaves, In: K. Korotkov (ed.), Measuring Energy Fields: State of the Science, Backbone Publishing, Fair Lawn, NJ, USA, 2004.