Serum Vitamin A Level Measurement in Slaughtered and Live Cattle. Using Multispectral Imaging

Transcription

1 Research Paper EAEF 3(2) : 42-46, 21 Serum Vitamin A Level Measurement in Slaughtered and Live Cattle Using Multispectral Imaging Noriko TAKAHASHI 1, Naoshi KONDO 2, Nguyen Quoc TUAN 1, Shoichi MANO 1, Tomoo SHIIGI 2, Hiroshi SHIMIZU 2, Moriyuki FUKUSHIMA 3, Fumiyuki IWAKI 3, Osamu WATANABE 3, Kazuyuki FUKUZONO 4, Mamoru NAKANO 5 Abstract The optimum wavelength for measuring pupil reflection in order to estimate serum vitamin A level in slaughtered and live cattle was investigated. A multispectrum camera was used to obtain the eye images. The highest coefficient of determination.76 between pupil reflection and serum vitamin A level was obtained at 5 nm. Since the photosensitive pigment rhodopsin has a maximum absorption of about 5 nm, the results suggest that 5 nm could be the optimum wavelength for measuring pupil reflection. [Keywords] beef quality, multispectrum camera, non-invasive, serum vitamin A I Introduction Serum vitamin A (V.A) is essential for cattle growth and is an important indicator of beef quality in Japan (Adachi et al., 1999). To produce high quality meat, serum V.A should be maintained at a low level, about 3 IU/dl, in cattle from 16 to 24 months of age. Since V.A deficiency (less than 3 IU/dl) can induce serious diseases in cattle, e.g., night blindness, diarrhea, xerophthalmia, convulsive seizures, and other infections (O Donoghue, 1955; Adachi et al., 1998; National Research Council, 2), it is essential to monitor serum V.A level carefully to maintain it above 3 IU/dl. Because the conventional blood assay used to measure serum V.A is time-consuming, expensive, and stressful to the cattle, a non-invasive method is needed. An unpublished study, in which 47-nm blue LED light was used with live cattle, showed that serum V.A level was related to pupil reflection. However, the optimum light wavelength to detect pupil reflection for estimating serum V.A level in live cattle is still unknown, and little is known about the relationship between pupil reflection and serum V.A in slaughtered cattle. The objective of the present study was to determine the optimum wavelength for measuring pupil reflection using a multispectrum camera in order to estimate serum V.A level in slaughtered and live cattle. II Materials Thirty live and eight slaughtered cattle were selected from cattle raised at the Hyogo Prefectural Hokubu Agricultural Institute. Live Japanese Black cattle and eye samples from slaughtered Japanese Black cattle were used to measure pupil reflection. 1. Live cattle eyes Pupil reflection was measured in the left eyes of thirty live cattle. A representative image of an eye from the live cattle is shown in Fig. 1(a). 2. Eyeballs Eight eyeballs were obtained from slaughtered cattle (Fig. 1 (b)). The eyeball samples were taken randomly from either right or left eyes. 3. Eyegrounds Eight eyegrounds were obtained by cutting the eyeballs and removing the glass body. The eyeground covered with retina includes a bright part known as the tapetum (Fig. 1 (c)) and a dark part (Fig. 1 (d)); the tapetum being symmetric for right and left eyes. The eyeground samples were taken randomly 1 Corresponding author: Graduate School of Agriculture, Kyoto University, Kitashirakawa-Oiwakecho, Sakyo-ku, Kyoto , Japan. norikot@kais.kyoto-u.ac.jp 2 JSAM Member, Graduate School of Agriculture, Kyoto University, Kitashirakawa-Oiwakecho, Sakyo-ku, Kyoto , Japan 3 Hyogo Prefectural Hokubu Agricultural Inst., Department of Beef Cattle Production, 123 Yasui, Wadayama, Asago, Hyogo , Japan 4 Seiwa Electric MFG Co., Ltd., Development Laboratory, 36 Terada-Shinike, Joyo, Kyoto , Japan 5 YPTECH Co., Ltd. Yuraku-cho Building, 1-1 Yuraku-cho 1-chome, Chiyoda-ku, Tokyo 1-6, Japan

2 TAKAHASHI, KONDO, TUAN, MANO, SHIIGI, SHIMIZU, FUKUSHIMA, IWAKI, WATANABE, 43 FUKUZONO, NAKANO : Serum Vitamin A Level Measurement in Slaughtered and Live Cattle Using Multispectral Imaging from either right or left eyes, and both bright and dark parts of the eyeground were examined. III Devices and methods 1. Experimental devices The multispectrum camera used in the present study, comprising a CCD camera (FC145, Takex, Japan) and tunable filter (VariSpec, Cambridge Research & Instrumentation, Inc., USA), is shown in Fig. 2. The spectrum distribution ranged from to 72 nm with a resolution of 1 nm. The camera condition was adjusted for capturing images of live and slaughtered cattle eyes, as shown in Table 1. Halogen lamps (12V5WAKW, Philips, Japan) were used to obtain a pupil reflection, at a light intensity of, lx for live cattle and 12, lx for slaughtered cattle. The spectrum distribution in transmittance of the tunable filters and lens, sensitivity of the CCD camera, relative energy of the halogen lamp, and total relative sensitivity of these devices are shown in Fig. 3. (a) (c) (b) (d) Fig. 1 Sample images of (a) live bovine eye; (b) eyeball; (c) and (d) eyeground. The eyeball and eyeground were obtained from slaughtered cattle. The eyeground includes a bright part called the tapetum (c) and a dark part (d). Relative transmittance (%) Sensitivity (%) Relative sensitivity (%) Table 1 Multispectrum camera condition Live cattle eye Eyeball* Eyeground* *obtained from slaughtered cattle Gain Offset Shutter speed (ms) (a) (b) (c) Relative energy (%) Fig. 2 Multispectrum camera used in the present study. Fig. 3 Spectrum distribution in transmittance of (a) tunable filters (solid line) and lens (dotted line), (b) sensitivity of CCD camera (solid line) and relative energy of halogen lamp (dotted line), and (c) total relative sensitivity of CCD camera, lens, tunable filters, and halogen lamp.

3 44 Engineering in Agriculture, Environment and Food Vol. 3, No. 2 (21) 2. Methods The images were processed using image processing software (WinROOF, Mitani Corporation, Japan) to extract the pupil area from the image and to measure the mean of gray scale (Figs. 4 and 5). First, the pupil area was detected. Second, the image was binarized with threshold 25 to eliminate the halation by lighting. Third, the image was inverted to obtain the mean of gray scale. Serum V.A level was simultaneously measured using a conventional blood assay test as the pupil reflection was measured by the camera. Both bright and dark parts of the eyeground were examined to determine the mean of gray scale for reflection measurement. The relationship between pupil reflection and serum V.A level was subsequently investigated. Pupil area detection Binarization (Threshold: 25) Inversion Histogram calculation gray scale from pupil reflection increased linearly with increasing serum V.A level (Fig. 7). Since the photosensitive pigment rhodopsin has a maximum absorption of about 5 nm (Kim et al., 21, 23), these results suggest that 5 nm could be the optimum wavelength for measuring pupil reflection Fig. 6 Coefficient of determination between pupil reflection and serum vitamin A level in slaughtered cattle at different wavelengths. 6 Fig. 4 Flow chart of image processing to determine the mean of gray scale. Meanofgrayscale R²= SerumVitaminAlevel(IU/dl) Fig. 7 Relationship between the mean of gray scale from pupil reflection and serum vitamin A level in slaughtered cattle at 5 nm. Fig. 5 Eye image processing using WinROOF software. The eye sample shown here was obtained from slaughtered cattle. IV Results and discussion 1. Eyeball from slaughtered cattle Serum V.A level in slaughtered cattle was 27 to 112 IU/dl. The coefficient of determination between pupil reflection and serum V.A level ranged from.21 to.76, the highest value occurring at the 5-nm wavelength (Fig. 6). The mean of 2. Eyeground from slaughtered cattle eye Both bright and dark parts of the eyeground were examined to determine the mean of gray scale for reflection measurement. No significant correlation between eyeground reflection at the bright part and serum V.A level was observed (Fig. 8), whereas a significant correlation was observed between eyeground reflection at the dark part and serum V.A level at 72 nm (Fig. 9). The coefficient of determination was.55 at 72 nm. Although there have been no reports on rhodopsin at this wavelength, this wavelength might be viable

4 TAKAHASHI, KONDO, TUAN, MANO, SHIIGI, SHIMIZU, FUKUSHIMA, IWAKI, WATANABE, 45 FUKUZONO, NAKANO : Serum Vitamin A Level Measurement in Slaughtered and Live Cattle Using Multispectral Imaging for serum V.A estimation using eyeground reflection at the dark part. No significant correlation between reflection of the entire eyeground and serum V.A level was observed (Fig. 1). This result suggests that it is essential to examine both bright and dark parts of the eyeground to determine the correlation between serum V.A level and pupil reflection Fig. 8 Coefficient of determination between eyeground reflection (bright part) and serum vitamin A level in slaughtered cattle at different wavelengths Fig. 9 Coefficient of determination between eyeground reflection (dark part) and serum vitamin A level in slaughtered cattle at different wavelengths Fig. 1 Coefficient of determination between whole eyeground reflection (bright and dark parts) and serum vitamin A level in slaughtered cattle at different wavelengths. 3. Live cattle eye Serum V.A level in live cattle was 25 to 13 IU/dl. The coefficient of determination between pupil reflection and serum V.A level was from to.18 (Fig. 11). No significant correlation between pupil reflection and serum V.A level was observed at 5 nm (Fig. 12). Since lighting decreased pupil area (Fig. 13), the light intensity used would not have been optimum for measuring pupil reflection in live cattle. Taking the effect of the pupil area on these data into consideration, no significant correlation between the mean of gray scale divided by pupil area, and serum V.A level, was observed (Fig. 14). It has been reported that the speed of pupil reflex to light is related to the level of serum V.A (Matsuda et al., 2); therefore, it is essential to consider what light intensity should be used to obtain a normal pupil unaffected by lighting Fig. 11 Coefficient of determination between pupil reflection and serum vitamin A level in live cattle at different wavelengths.

5 46 Engineering in Agriculture, Environment and Food Vol. 3, No. 2 (21) Meanofgrayscale R²= Serum VitaminA level(iu/dl) Fig. 12 Relationship between the mean of gray scale from pupil reflection and serum vitamin A level in live cattle at 5 nm. Meanofgrayscale/pupilarea R²=.164 Pupil Fig. 13 Example of pupil image using a multispectrum camera in live cattle Serum VitaminA level(iu/dl) Fig. 14 Relationship between mean of gray scale divided by pupil area, and serum vitamin A level in live cattle at 5 nm. V Conclusions A linear relationship between pupil reflection and serum V.A level was shown, the highest coefficient of determination (.76) occurring at 5 nm in slaughtered cattle. No significant correlation between eyeground reflection at the bright part and serum V.A level was observed, whereas a significant correlation between serum V.A level and eyeground reflection at the dark part was observed at 72 nm. However, the significance of this wavelength should be investigated further. For live cattle, no significant correlation between pupil reflection and serum V.A level was observed at the 5-nm wavelength. Therefore, the appropriate light intensity for obtaining a normal pupil unaffected by lighting must be considered in future studies. Acknowledgements This study was supported by the Bio-oriented Technology Research Advancement Institution (28-21). References Adachi, K., H. Kawamoto, Y. Komura, Y. Yamamoto, A. Arikawa, A. Tsujii, M. Adachi, T. Onumaru, K. Ohwada Relationship between serum biochemical values and marbling scores in Japanese black steers. Journal of Veterinary Medicine Science 61: Adachi, K., K. Fukumoto, Y. Komura, N. Katsura, A. Arikawa, A. Tsuji, T. Onumaru, Significant decrease of serum vitamin A level in Japanese black beef steers after introduction to a farm. Journal of Veterinary Medicine Science 6: Kim, J. E., M. J. Tauber, R. A. Mathies, 21. Wavelength dependent Cis-Trans Isomerization in Vision. Biochemistry : Kim, J. E., M. J. Tauber, R. A. Mathies, 23. Analysis of the mode-specific excited-state energy distribution and wavelength-dependent photoreaction quantum yield in rhodopsin. Biophysical Journal 84: Matsuda, K., A. Watanabe, T. Ichijo, S. Yashima, K. Ujiie, A. Kawana, The relationship between blood vitamin A concentration and oculopupillary reflex in Japanese black fattening cattle. Journal of Veterinary Clinic 47: National research council, 2. Nutrient requirements of beef cattle. Seventh revised edition National academy press, United States. O donoghue, J. G., Vitamin A deficiency in beef cattle. Canadian Journal of Comparative Medicine 12: (Received: 29. July. 29, Accepted: 22. January. 21)