Author Index
A
Achmad Munir
C4-1 (Pp.298-302);
C4-2 (Pp.303-306)
Aciek Ida Wuryandari
B1-1 (Pp.97-102)
Adang Suwandi Achmad
A1-1 (Pp.21-25);
SS1 (Pp.1-6);
SS2 (Pp.7-10);
SS3 (Pp.11-15);
SS4 (Pp.16-20);
B2-2 (Pp.137-142)
Addin Suwastono
B4-3 (Pp.188-192)
Ade Ramdan
B3-3 (Pp.158-162)
Adi Soeprijanto
B1-2 (Pp.103-108)
Adi Sucipto
C3-1 (Pp.272-277)
Adit Kurniawan
A2-5 (Pp.68-72)
Agung Nuza Dwiputra
D5-2 (Pp.365-370)
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)
Astri Maria
A2-3 (Pp.57-62)
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)
Djohar Syamsi
B1-1 (Pp.97-102);
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);
Febry Ramos Sinaga
D5-4 (Pp.376-379)
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
C4-2 (Pp.303-306)
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);
Kenta Omiya
C3-4 (Pp.287-292)
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)
Mochamad Fahri
B3-1 (Pp.148-152)
Mochamad Hariadi
B3-1 (Pp.148-152)
Mochamad Irwan Nari
A3-4 (Pp.88-91)
Mochammad Alif Ramadhan
B3-5 (Pp.168-173)
Moh Amanta KS Lubis
D4-1 (Pp.343-346)
Moh Hasbi Assidiqi
B1-5 (Pp.121-126)
Mohamad Nasyir Tamara
A1-3 (Pp.30-34)
Mohammad Nuh
A1-4 (Pp.35-40)
Monang Kevin Napitupulu
B4-4 (Pp.193-198)
MS Kasbe
B1-4 (Pp.115-120)
MS Kasbe
B3-6 (Pp.174-178)
Muhamad Amin Abdul Wahab
A3-2 (Pp.79-84)
Muhamad Komarudin
D5-4 (Pp.376-379)
Muhammad Amin Sulthoni
D5-1 (Pp.360-364)
Muhammad Ammar Wibisono
C4-2 (Pp.303-306)
Muhammad Arief Ma'Ruf Nasution
D5-2 (Pp.365-370)
Muhammad Arsyad
A2-2 (Pp.51-56)
Nanna Suryana
C3-5 (Pp.293-297)
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
B4-4 (Pp.193-198)
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)
Rizqia Cahyaningtiyas
A1-2 (Pp.26-29)
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);
B1-5 (Pp.121-126)
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);
Syifaul Fuada
A2-4 (Pp.63-67);
B1-6 (Pp.127-131);
D5-3 (Pp.371-375)
T
Takeshi Shima
B4-2 (Pp.184-187)
Tatag Budiardi
B3-4 (Pp.163-167)
Taufik Ibnu Salim
B5-2 (Pp.208-213)
TH Mujawar
B1-4 (Pp.115-120)
B3-6 (Pp.174-178)
Thanh Duc Ngo
C1-4 (Pp.242-247)
Thoriq Satriya
C1-6 (Pp.253-257)
Tien Do
C1-4 (Pp.242-247)
Tiper Uniplaita
C1-3 (Pp.236-241)
Titin Yulianti
C1-5 (Pp.248-252)
Trio Adiono
A2-4 (Pp.63-67);
B1-6 (Pp.127-131);
C5-1 (Pp.315-319);
SS2 (Pp.7-10)
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
Meat Quality Classification Based on Color Intensity
Measurement Method
Titin Yulianti1,a, Afri Yudamson1,b, Hery Dian
Septama1,c, Sri Ratna Sulistiyanti 1,d, F.X.Arinto
Setiawan1,e
1Department of Electrical Engineering,
University of Lampung, Bandar Lampung, Indonesia
atitin.yulianti@eng.unila.ac.id,
bafri.yudamson@eng.unila.ac.id, chery@eng.unila.ac.id, dsriratnasulistiyanti@gmail.com, efx.arinto@eng.unila.ac.id,
Mareli Telaumbanua2,f
2Department of Agriculture Engineering,
University of Lampung, Bandar Lampung, Indonesia
fmareli.telaumbanua@fp.unila.ac.id
Abstract— 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 isdefined, 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. INTRODUCTION
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].
The beef image can be extracted by using computer analysis features. Through analysis process and classification
2016 International Symposium on Electronics and Smart Devices (ISESD) November 29-30, 2016
using computational algorithms, the meat quality information 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 percentage of color intensity average from R, G, and B channel. The optimal feature is gotten based on the percentage values.
II. RELATED WORKS
The fresh and defective meat identification can be performed by laboratory tests. However, the access is limited only by food quality associated institutions. Guzek et al [5] 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 spectroscopy near distance and computer-based image analysis. 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 space and histogram. Red Green Blue (RGB) and Hue Sturation Intensity (HSI) color space were used in the research. 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]. The color and multispectral image texture features were used on beef tenderness prediction [7].
Yuristiawan [8] developed an aplication for local beef freshness level detection using feature extraction of color statistical approach.
III. MATERIAL AND METHOD
A. Data Preparation
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 beef, the number of data are 30 images.
B. Approach
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. 1 Flowchart of approach image processing
The first step is image preprocessing by cropping the image to get the RoI (Region of Interest) and eliminating the image label and the background. In this research the RoI of image is the beef as the object. The example of the cropped image is shown in Fig 2.
Fig. 2 Example of beef images after 4 hours (first row) and beef images after 16 hours (second row) that have been cropped
Measuring Intensity average value of R, G, and B channel
Start
RGB Image
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.
1 1
Then the percentage of the separate color intensity average of each channel is measured by using equation (4), (5), and (6).
%R R therefore the optimal feature is obtained.
IV. RESULT AND DISCUSSION
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 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. 3 The alteration of percentages of intensity value consecutively every 4 hours on beef sample 1
Fig. 4 The alteration of percentages of intensity value consecutively every 4 hours on beef sample 2
Fig. 5 The alteration of percentages of intensity valueconsecutively every 4 hours on beef sample 3
Fig. 6 The alteration of percentages of intensity value consecutively every 4 hours on beef sample 4
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.
The minimum, maximum, and average value of the percentages of intensity average are tabulated in Table 1.
TABLE I. THE COMPARISON OF PERCENTAGES OF INTENSITY AVERAGE VALUE
Beef Fresh Defective
%R Min 56.38 37.76 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.
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 and defective beef. The fresh beef has percentage of R channel intensity average value range from 56.38% to 66.33%, whereas
the percentage value range for defective beef is 37.76% to 51.71%.
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.
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%.
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.
V. CONCLUSION
The measurements of the color intensity average on the R, G, and B channel of the beef image is presented in this paper. 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.
VI. FUTURE WORKS
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.
ACKNOWLEDGMENT
The authors would like to thank Integrated Control System (ICS) Riset Group of Electrical Engineering, University of Lampung and also thank to LPPM for providing financial support through DIPA PNBP Faculty of Engineering.
REFERENCES
[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 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
ed. Upper Saddle River, New Jersey: Prentice Hall, 2008.
[4] T. Yulianti, "No-reference Retinal Image Quality Assessment Method Development Based on Feature Extraction (in bahasa: Pengembangan Metode Penilaian Kualitas Citra Retina Tanpa Menggunakan Citra Referensi Berbasis Ekstraksi Fitur)," Electrical Engineering, Universitas Gadjah Mada, Yogyakarta, Indonesia, 2015.
[5] D. Guzek, D. Glabska, E. Pogorzelska, G. Pogorzelski, and A. Wierzbicka, "Instrumental Texture Measurement of Meat in A Laboratory Research and on A Production Line," Advances in
Sience and Technology Journal, vol. 7, September 2013.
[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.
[7] X. Sun, K. J. Chen, K. R. Maddock-Carlin, V. L. Anderson, A. N. Lepper, C. A. Schwartz, et al., "Predicting beef tenderness
using color and multispectral image texture feature," Meat
Science Journal, vol. 92, pp. 386-393, December 2012.
[8] D. Yuristiawan, F. Z. Rahmanti, and H. A. Santoso, "Application for Beef freshness level detection using color extraction features with statistics Approach (in bahasa : Aplikasi Pendeteksi Tingkat Kesegaran Daging Sapi Lokal enggunakan Ekstraksi Fitur Warna dengan Pendekatan Statistika," Riptek,
vol. 9, pp. 9-11, 2015.
