Meat Quality Classification Based on Color Intensity Measurement Method(1)
IEEE Catalog Number: CFP16J03-ART
Author Index
Achmad Munir C4-1 (Pp.298-302); C4-2 (Pp.303-306) B1-1 (Pp.97-102)
A
Aciek Ida Wuryandari A1-1 (Pp.21-25);
Adang Suwandi Achmad SS1 (Pp.1-6); SS2 (Pp.7-10); SS3 (Pp.11-15); SS4 (Pp.16-20); B2-2 (Pp.137-142) B4-3 (Pp.188-192)
Addin Suwastono B3-3 (Pp.158-162)
Ade Ramdan B1-2 (Pp.103-108)
Adi Soeprijanto C3-1 (Pp.272-277)
Adi Sucipto A2-5 (Pp.68-72)
Adit Kurniawan C1-5 (Pp.248-252)
Afri Yudamson
Agung Nuza Dwiputra Agus Bejo
B4-3 (Pp.188-192); C2-2 (Pp.264-267)
Ahmad Zainudin A2-2 (Pp.51-56)
Akbari Indra Basuki B1-1 (Pp.97-102)
Andri Fachrur Rozie B3-3 (Pp.158-162)
Andriyan Bayu Suksmono C5-2 (Pp.320-325)
Andriyan Suksmono C1-3 (Pp.236-241)
Andryan Bagoes Noegroho A2-2 (Pp.51-56)
Angga Pratama Putra D5-3 (Pp.371-375)
Angga Putra A2-4 (Pp.63-67)
Annisa Maulidary Muthiah C5-3 (Pp.326-330)
Ardhi Maarik B3-4 (Pp.163-167)
Arief Darmawan B4-3 (Pp.188-192)
Arif Sasongko B1-3 (Pp.109-114)
Arwin Datumaya Wahyudi Sumari SS1 (Pp.1-6); SS2 (Pp.7-10) C2-3 (Pp.268-271)
Astri Maria
B Bagas Mardiasyah Prakoso A2-2 (Pp.51-56) Bagas Prima Anugerah
D5-1 (Pp.360-364) Baharuddin Aziz
A3-1 (Pp.73-78) Bambang Anggoro
SS3 (Pp.11-15); B2-2 (Pp.137-142)
Barokatun Hasanah C4-1 (Pp.298-302)
Bijay Kumar Sahoo B5-4 (Pp.219-224)
Bima Sahbani B4-4 (Pp.193-198)
C Camallil Omar A3-2 (Pp.79-84) Catherine Olivia Sereati
SS2 (Pp.7-10) Chembian WT
C3-2 (Pp.278-282)
D Danny M Gandana B5-3 (Pp.214-218) Dedy Rahman Wijaya
C5-5 (Pp.337-342) A3-3 (Pp.85-87) Djohar Syamsi B3-3 (Pp.158-162)
Donny Danudirdjo C1-3 (Pp.236-241); C5-2 (Pp.320-325)
Dwi Nugroho Hari Wicaksono A1-3 (Pp.30-34)
E Efy Yosrita A1-2 (Pp.26-29) Eko Tjipto Rahardjo
D4-1 (Pp.343-346) Elvayandri
A3-1 (Pp.73-78) Enggar Fransiska Dwi Widyatama
C5-3 (Pp.326-330) Enny Zulaika
C5-5 (Pp.337-342)
F Fadhli Dzil Ikram C5-1 (Pp.315-319) Faisal Ardhy
E1-2 (Pp.386-393) Fajar Adiatmoko
B1-2 (Pp.103-108) Farkhad Ihsan Hariadi
A1-1 (Pp.21-25); B4-1 (Pp.179-183); E1-2 (Pp.386-393) Febry Ramos Sinaga Fitri Yuli Zulkifli
D4-1 (Pp.343-346) FX Arinto Setiawan
C1-5 (Pp.248-252)
G Goutam Mohanty B5-4 (Pp.219-224) Grasia Meliolla
B4-1 (Pp.179-183)
H Habibullah Akbar C3-5 (Pp.293-297) Hajiar Yuliana
A2-5 (Pp.68-72) Hamdan Prakoso
C3-3 (Pp.283-286) Harjito Bambang
B2-3 (Pp.143-147) Haruna Aimi
C2-1 (Pp.258-263) Helmy Fitriawan
D5-4 (Pp.376-379) Hendro Widodo
B1-2 (Pp.103-108) Heri Prasetyo
B2-3 (Pp.143-147) Hilman Syaeful Alam
B5-2 (Pp.208-213) Hiroshi Ochi
A2-3 (Pp.57-62) I Iskandar
I Wayan Sudiarta A3-3 (Pp.85-87)
Igi Ardiyanto C1-6 (Pp.253-257)
Ii Munadhif B1-2 (Pp.103-108)
Ilman Himawan Kusumah A3-4 (Pp.88-91)
Iskandar D4-3 (Pp.355-359)
Ismail Ariffin A3-2 (Pp.79-84)
J Jamil Akhtar B1-4 (Pp.115-120); B5-4 (Pp.219-224) Joko Suryana
A2-1 (Pp.46-50) Juhana Jaafar
A1-5 (Pp.41-45)
K Karel Octavianus B2-2 (Pp.137-142) Karel Octavianus Bachri
SS3 (Pp.11-15) Kenji Suyama
C1-2 (Pp.230-235); C2-1 (Pp.258-263); C3-4 (Pp.287-292) Kenta Omiya Khilda Afifah
B1-6 (Pp.127-131) Khoirul Anwar
D4-2 (Pp.347-354) Kiewlamphone Souvanlit
C2-2 (Pp.264-267) Kurnia Adi Nugroho
B4-1 (Pp.179-183)
L Lilik Subiyanto B1-2 (Pp.103-108) LP Deshmukh
B1-4 (Pp.115-120) B3-6 (Pp.174-178)
M Mahendra Drajat Adhinata B1-3 (Pp.109-114) Marcelinus Henry Menori
B2-1 (Pp.132-136) Mareli Telaumbanua
C1-5 (Pp.248-252) Mario Tressa Juzar
C1-1 (Pp.225-229) Mat Syai’in
C5-4 (Pp.331-336); B1-2 (Pp.103-108)
Maulana Yusuf Fathany B1-6 (Pp.127-131) A1-5 (Pp.41-45)
Mochamad Fahri Mochamad Hariadi
B1-4 (Pp.115-120) MS Kasbe
D5-2 (Pp.365-370) Muhammad Arsyad
C4-2 (Pp.303-306) Muhammad Arief Ma'Ruf Nasution
D5-1 (Pp.360-364) Muhammad Ammar Wibisono
D5-4 (Pp.376-379) Muhammad Amin Sulthoni
A3-2 (Pp.79-84) Muhamad Komarudin
B3-6 (Pp.174-178) Muhamad Amin Abdul Wahab
B4-4 (Pp.193-198) MS Kasbe
B3-1 (Pp.148-152) Mochamad Irwan Nari
A1-4 (Pp.35-40) Monang Kevin Napitupulu
A1-3 (Pp.30-34) Mohammad Nuh
B1-5 (Pp.121-126) Mohamad Nasyir Tamara
D4-1 (Pp.343-346) Moh Hasbi Assidiqi
B3-5 (Pp.168-173) Moh Amanta KS Lubis
A3-4 (Pp.88-91) Mochammad Alif Ramadhan
A2-2 (Pp.51-56) Nanna Suryana Naoki Shinohara
C1-2 (Pp.230-235) Nasril
B5-3 (Pp.214-218) Ndaru Anggit Wicaksono
D5-1 (Pp.360-364) Nevi Faradina
D4-3 (Pp.355-359) Ngoc-Bao Nguyen
C1-4 (Pp.242-247) Nico Surantha
A2-3 (Pp.57-62) Nicodimus Retdian
B4-2 (Pp.184-187) NN Maldar
B1-4 (Pp.115-120) Noorman Rinanto
B1-2 (Pp.103-108); B3-5 (Pp.168-173); C5-4 (Pp.331-336)
Novi Prihatiningrum B1-3 (Pp.109-114)
O Octarina Nur Samijayani C5-3 (Pp.326-330) Oka Mahendra
B3-3 (Pp.158-162) Oktanto Dedi Winarko
C5-3 (Pp.326-330) Panji Ramadhan Pristy Ar Nurisysyifak
A2-2 (Pp.51-56)
R Rachmad Vidya Wicaksana Putra B1-6 (Pp.127-131) Radhian Ferel Armansyah
C5-1 (Pp.315-319) Rahmadina Alamsyah
B5-3 (Pp.214-218) Raja Fathurrahim Akmaludin
A3-3 (Pp.85-87) Rakhmat Arianto
A1-2 (Pp.26-29) Rengga Yanuar Putra
B3-2 (Pp.153-157) Retno Tri Wahyuni
A3-5 (Pp.92-96) Ricky Disastra
B1-3 (Pp.109-114) Ridi Ferdiana
C3-3 (Pp.283-286) Riko Hasiando Goknipasu Nainggolan
D5-2 (Pp.365-370) Rinaldi Munir
B2-1 (Pp.132-136) Rinaldi Munir
C1-1 (Pp.225-229) Risanuri Hidayat
C2-2 (Pp.264-267) Riyanarto Sarno
C5-5 (Pp.337-342) C5-2 (Pp.320-325) Rizqia Cahyaningtiyas Rubita Sudirman
A1-5 (Pp.41-45) Rubita Sudirman
A3-2 (Pp.79-84) Rudy Hartanto
C3-3 (Pp.283-286) Ryan Adhitya
B1-2 (Pp.103-108); C5-4 (Pp.331-336)
Ryan Yudha Adhitya B3-2 (Pp.153-157); B3-5 (Pp.168-173)
S Salih Ergun B4-5 (Pp.199-202) Samudra Arrachman
B1-2 (Pp.103-108) Sarwono Sutikno
SS4 (Pp.16-20) Seetharaman Krishnamoorthy
C3-2 (Pp.278-282) Sena Sukmananda Suprapto
A3-4 (Pp.88-91) Septafiansyah Dwi Putra
SS4 (Pp.16-20) S-Erlyane Rosli
A1-5 (Pp.41-45) Son Kuswadi
A1-3 (Pp.30-34); A1-4 (Pp.35-40); Sri Ratna Sulistiyanti C1-5 (Pp.248-252)
Sri Wahjuni B3-4 (Pp.163-167)
Sryang Sarena B1-2 (Pp.103-108)
Sryang Tera Sarena C5-4 (Pp.331-336)
SS Mule B1-4 (Pp.115-120); B3-6 (Pp.174-178);
S-Syakiylla S-Daud A1-5 (Pp.41-45)
Suci Rahmatia C5-3 (Pp.326-330)
Sunu Wibirama C1-6 (Pp.253-257)
Supeno Mardi Susiki B3-1 (Pp.148-152)
Surya Ramadhan A1-1 (Pp.21-25)
Susi Juniastuti B3-1 (Pp.148-152)
Swizya Satira Nolika C5-1 (Pp.315-319)
Syaiful Alam D5-4 (Pp.376-379)
Syamsiar Kautsar B1-2 (Pp.103-108)
Syamsiar Kautsar B3-2 (Pp.153-157); B3-5 (Pp.168-173)
Syifaul Fuada B1-6 (Pp.127-131); D5-3 (Pp.371-375)
B4-2 (Pp.184-187) Takeshi Shima
T
B3-4 (Pp.163-167) Tatag Budiardi
B5-2 (Pp.208-213) Taufik Ibnu Salim
B1-4 (Pp.115-120) TH Mujawar
B3-6 (Pp.174-178) C1-4 (Pp.242-247)
Thanh Duc Ngo C1-6 (Pp.253-257)
Thoriq Satriya C1-4 (Pp.242-247)
Tien Do C1-3 (Pp.236-241)
Tiper Uniplaita C1-5 (Pp.248-252)
Titin Yulianti A2-4 (Pp.63-67);
Trio Adiono B1-6 (Pp.127-131); C5-1 (Pp.315-319); D5-3 (Pp.371-375);
Triya Haiyunnisa B5-2 (Pp.208-213)
Tuppak Bobby Vorlen Sagala A2-1 (Pp.46-50)
Tutun Juhana C4-3 (Pp.307-310)
V Vu-Hoang Nguyen C1-4 (Pp.242-247)
W Wahyul Amin Syafei C4-4 (Pp.311-314)
Y Yoanes Bandung C3-1 (Pp.272-277); E1-1 (Pp.380-385) Yuhei Nagao
A2-3 (Pp.57-62) Yulian Aska
A2-4 (Pp.63-67); D5-3 (Pp.371-375)
Yusmar Palapa Wijaya A3-5 (Pp.92-96)
Z Zulhamdi Koto E1-1 (Pp.380-385)
Meat Quality Classification Based on Color Intensity
Measurement Method
Titin Yulianti 1,a , Afri Yudamson 1,b , Hery Dian Septama 1,c , Sri Ratna Sulistiyanti 1,d , F.X.Arinto Setiawan 1,e 1 Department of Electrical Engineering, University of Lampung, Bandar Lampung, Indonesia a titin.yulianti@eng.unila.ac.id, b afri.yudamson@eng.unila.ac.id, c hery@eng.unila.ac.id, d sriratnasulistiyanti@gmail.com, e fx.arinto@eng.unila.ac.id, Mareli Telaumbanua 2,f 2 Department of Agriculture Engineering, University of Lampung, Bandar Lampung, Indonesia f mareli.telaumbanua@fp.unila.ac.idAbstract— The fresh and defective beef identification by consumers is subjectively through visual observation. However, identifying beef quality manually has disadvantage, there is human visual limitations, differences in human perception in assessing the quality of an object, and ability of each individual knowledge are different. Therefore, we need a technological device that can be applied to identify the quality of beef that can be used by people. The aim of this research is measuring the percentage of color intensity average from R, G, and B channel. The fresh and defective beef is identified using feature of the beef image. That feature is percentages of intensity average value from R (red), G (green), and B (blue) channel. The optimal feature is gotten based on the percentage values. The feature is gotten by using image processing method. The percentage of R channel intensity average value is defined, which can be used to classify the fresh and defective beef . The percentage of R channel intensity is consecutively decrease on every 4 hours. It is shown on each beef sample. The R channel of the fresh image has higher percentage of intensity average value than the defective beef. The fresh beef has 56.38% to 66.33% of the R channel intensity average. whereas the defective beef has 37.76% to 51.71% of the R channel intensity. Keywords—percentage of intensity average, beef quality classification, image pocessing.
I. I NTRODUCTION Data from National Survey of Social Economic in Indonesia (SUSENAS) year 2014 showed that Indonesian consumption of beef is only 2.08 kg / capita / year. This number is lower than beef consumption in other developed countries. In general, the Indonesian people consume beef mostly at celebrations and religious holidays [1].
The potential of cattle breeding development for meat demand in Indonesia is very large. The availability of land, labour, and the capacity of natural resources is abundant. Moreover, the government support, making cattle breeding sector in Indonesia become potential.
However, Indonesia still not be able to fulfill beef stock for nationaldemand. Therefore, Indonesia is depending on import to overcome the situation.
The location of cattle farm in Indonesia is also not evenly distributed in each province. This resulted in a lack of availability meat and an increase price of meat in an area with a great level of meat consumption. The cattle production centers in Indonesia are in East Java province that is equal to 21.09% of beef production throughout Indonesia, while the province of Lampung produce only 2.44% of national beef production.
Based on Information System for Agriculture in 2015, the development of beef prices at the consumer level from 1983 to 2015 has fluctuated and tended to increase. During these periods, the price of beef at the consumer level rose by 13.21% per year. Beef prices last five-year period (2011-2015) tend to increase Rp.69.641 to Rp.104.326 [1].
The high price of beef cause to a few unfair traders take action to mix the fresh beef with defective beef to obtain greater profits. Thus, the problem of defective beef sales in the market are still happening. The inspection and investigation conducted by government has not been able to guarantee that traders did not sells defective beef. Therefore, consumers need the ability to identify beef quality, before buying it.
Until now, fresh and defective beef identification method by consumers is subjectively through visual observation [2]. However, identifying beef manually has disadvantage, there is human visual limitations, differences of human perception on assessing the quality of an object, and ability of each individual knowledge are different [2-4]. Therefore, we need a technological device that can be applied to identify the quality of beef that can be used by people. The first step in research that starts from develop of a method for identifying fresh beef and defective ones. The method is used based on image processing, because the image of meat are able to represent its quality [2].
2016 International Symposium on Electronics and Smart Devices (ISESD) November 29-30, 2016 using computational algorithms, the meat quality information Start can be obtained.
This work is focused on identify of beef quality and clasify it as fresh and defective. The aim of this study is measuring the RGB Image percentage of color intensity average from R, G, and B channel. The optimal feature is gotten based on the percentage values.
R, G, and B channel ELATED ORKS extracting
II. R W The fresh and defective meat identification can be performed by laboratory tests. However, the access is limited
Measuring Intensity average only by food quality associated institutions. Guzek et al [5] value of R, G, and B channel studied the appropriate way to analyze and develop method to identify meat quality outside the laboratory. The results of this study is a method of meat identification using infrared
Measuring percentage of R, spectroscopy near distance and computer-based image analysis.
G, and B channel The research related to the identification of meat has been conducted by several researchers. Nai chian et al [2] classified the meat freshness degree using texture and the change of color
Percentage of space and histogram. Red Green Blue (RGB) and Hue intensity Sturation Intensity (HSI) color space were used in the research. average value
Mean value and mean interval value of color space were used in classification. The other research investigated that the color change in foal meat can vary after thawing out in relation to slaughtering age of the horses and to the post thawing time [6]. End The color and multispectral image texture features were used on beef tenderness prediction [7]. Fig. 1 Flowchart of approach image processing Yuristiawan [8] developed an aplication for local beef
The first step is image preprocessing by cropping the freshness level detection using feature extraction of color image to get the RoI (Region of Interest) and eliminating the statistical approach. image label and the background. In this research the RoI of image is the beef as the object. The example of the cropped ATERIAL AND M ETHOD image is shown in Fig 2. A. Data Preparation
III. M The tenderloin beef that commonly used for steak is used as the sample. Furthermore, the beef is sliced crosswise as five pieces and placed on the plate. The smartphone’s camera with resolution of 5MP is used to capture the beef images. Since the resolution is commonly used on smartphone and as the minimum resolution of smartphone’s camera today. We assume when using camera with resolution of 5MP can identify the beef quality, it is mean with higher resolution the beef quality can be identified easier. The images are taken every 4 hours consecutively in 24 hours. Since there are 5 samples of B. Approach beef, the number of data are 30 images.
In this research , fresh and defective beef is identified using feature of the beef image. That feature is percentages of intensity average value from R (red), G (green), and B (blue) channel. The feature is gotten by using image processing method. The steps of the image processing is shown in Fig.1. Fig. 2 Example of beef images after 4 hours (first row) and beef images after 16 hours (second row) that have been cropped
Since the image is in RGB color, the channels can be extracted. The measurement of the separate color intensity average is done in each channel by using the equations below. M − 1 N − 1 R I m n = [ , ] (1)
1 R M N × m = n = M − N − 1 1 G I m n
1 = G [ , ] (2) M N m = n = × M − − 1 N 1 B I m n (3) = [ , ]
1 B M N = = × m n R , I G 4 hours on beef sample 1 Fig. 3 The alteration of percentages of intensity value consecutively every Where M and N are the length and width of image. I , I B and are intensity of R, G, and B channel. [m, n] is coordinate of the each pixel.
Then the percentage of the separate color intensity average of each channel is measured by using equation (4), (5), and (6). R R % = (4) G R G B G + + % (5) B = R G B + + B Fig. 4 The alteration of percentages of intensity value consecutively every % (6) = R G B + + 4 hours on beef sample 2 The number of the percentage needs to be selected, therefore the optimal feature is obtained. ESULT AND ISCUSSION
IV. R D During beef observation by taken image of beef consecutively every 4 hours, the beef its self has decomposed. The decomposed process can be observed visually based on the beef color of image. However, the color change is subjective and has not been measurable yet. It means that identifying of beef freshness is depend on observer experience. The freshness level of beef can be identified by using image processing 4 hours on beef sample 3 Fig. 5 The alteration of percentages of intensity value consecutively every method conducted in this research.
The result of this research is shown in graph. Fig.3- Fig.7 show the alteration percentages of intensity value of R, G, and B channel consecutively every 4 hours on each beef sample.
Fig. 7 The alteration of percentages of intensity value consecutively every 4 hours on beef sample 5 The results show that the RGB channel intensity has a common pattern. The percentage of R channel intensity value are consecutively decrease every 4 hours and more significantly decrease among 12 and 16 hours. Whereas the percentage of B channel intensity value mostly increase every 4 hours and more significantly increase among 12 and 16 hours. However, the percentage of G channel intensity value did not show alteration significantly. Therefore we assumed that we can clasify the meat quality by using the color intensity measurements. The results show that first 12 hours may clasified as fresh meat and after 12 hours as defective meat.
22.85
R EFERENCES [1] R. Suryani, "Agricultural Commodities Outlook: beef livestock subsector (in bahasa : Outlook komoditas pertanian subsektor peternakan daging sapi)." Sekretariat Jenderal, Kementerian Pertanian, Indonesia2015. [2] V. N. Chian, F. S. A. Saad, M.F.Ibrahim, S. Sudin, A. Zakaria, and A. Y. M. Shakaff, "Meat Color Recognition and
Lampung and also thank to LPPM for providing financial support through DIPA PNBP Faculty of Engineering.
A CKNOWLEDGMENT The authors would like to thank Integrated Control System (ICS) Riset Group of Electrical Engineering, University of
VI. F UTURE W ORKS The R, G, and B channel pattern in this paper for beef quality classification may be used to classified another meat that have closely characteristics with beef i.e red color. The others should observed as future works to find the R, G, and B channel pattern.
The percentage of R channel intensity is consecutively decrease on every 4 hours and more significantly decrease among 12 and 16 hours.. It is shown on each beef sample. The percentage of color intensity average of each channel is also measured. The R channel of the fresh image has higher percentage of intensity average value than the defective beef. The fresh beef has 56.38% to 66.33% of the R channel intensity average. whereas the defective beef has 37.76% to 51.71% of the R channel intensity. Therefore, the percentage of the color intensity average of the Red channel on beef image can be used as the feature to identify the fresh and defective beef.
V. C ONCLUSION The measurements of the color intensity average on the R, G, and B channel of the beef image is presented in this paper.
It means that the proposed method is successfully obtain the optimal feature. The percentage of R channel intensity average value is defined, which can be used to separate the fresh and defective beef.
Then, the maximum value of the percentage on G channel of fresh beef and the minimum value of percentage on G channel of defective beef have same value, 22.85%.
On the B channel, the minimum percentage value of fresh beef is 18.05% while the maximum value of the defective beef is 22.96%. The two values are closely intersect, thus it can’t be used as the feature.
The minimum value of the percentage on R channel of fresh beef is 56.38%, therefore the maximum value of percentage on R channel of defective beef is 51.71%. It is mean that it is can be used as the feature to identify the fresh the percentage value range for defective beef is 37.76% to 51.71%.
28.14 The percentages of R channel intensity average value on the fresh beef is higher than the defective beef. Whereas the percentages of B channel intensity average value on the fresh beef is mostly lower than the defective beef. However, the percentages of G channel intensity average value on the fresh and defective beef are fluctuating.
17.15
32.25 Average
22.85 Max
The minimum, maximum, and average value of the percentages of intensity average are tabulated in Table 1. TABLE I. T HE COMPARISON OF PERCENTAGES OF INTENSITY AVERAGE VALUE Beef Fresh Defective %R Min
11.61
26.20 %B Min
21.88
30.36 Average
25.31
22.96 Max
18.05
46.66 %G Min
60.97
51.71 Average
66.57
37.76 Max
56.38
Classification Based on Color using NIR/VIS Camera," presented at the 8 th MUCET, Melaka, Malaysia, 2014. [3] R. C. Gonzales and R. E. Woods, Digital Image Processing, 3rd
[4] T. Yulianti, "No-reference Retinal Image Quality Assessment [7] X. Sun, K. J. Chen, K. R. Maddock-Carlin, V. L. Anderson, A.
Method Development Based on Feature Extraction (in bahasa: N. Lepper, C. A. Schwartz, et al., "Predicting beef tenderness Pengembangan Metode Penilaian Kualitas Citra Retina Tanpa using color and multispectral image texture feature," Meat Science Journal, Menggunakan Citra Referensi Berbasis Ekstraksi Fitur)," vol. 92, pp. 386-393, December 2012.
Electrical Engineering, Universitas Gadjah Mada, Yogyakarta, [8] D. Yuristiawan, F. Z. Rahmanti, and H. A. Santoso, Indonesia, 2015.
"Application for Beef freshness level detection using color
[5] D. Guzek, D. Glabska, E. Pogorzelska, G. Pogorzelski, and A. extraction features with statistics Approach (in bahasa : Aplikasi
Wierzbicka, "Instrumental Texture Measurement of Meat in A Pendeteksi Tingkat Kesegaran Daging Sapi Lokal enggunakan Sience and Technology Journal, Laboratory Research and on A Production Line," Advances in Ekstraksi Fitur Warna dengan Pendekatan Statistika," Riptek, vol. 7, September 2013. vol. 9, pp. 9-11, 2015.[6] P. D. Palo, A. Maggiolino, P. Centoducati, and A. Tateo, "Colour Changes in Meat of Foals as Affected by Slaughtering Age and Post-thawing Time," Asian-Aust. J. Anim. Sci., vol. 25, pp. 1775-1779, December 2012.