Discrimination of civet coffee using near infrared spectroscopy and artificial neural network

Abstract

Civet coffee refers to the coffee that includes partlydigested coffee cherries eaten and defecated by the palm civet cat (Paradoxurus hermaphrodites). Civet cats seek out and eat only the ripest, reddish coffee cherries. The coffee cherry fruit is sweet and is completely digested, but the beans are excreted in their feces. It is one of the rarest and priciest coffees in the world [1]. At present, there is still no accepted standard in determining the authenticity of the civet coffee [2]. Due to its high price, the traditional method of collecting feces from the wild civets has given way to a farming method where civets are kept in cages and force-fed with the cherries. Most of the previous studies done in discriminating civet coffee from ordinary coffee beans used a destructive method of sampling, preparation wherein coffee beans were grounded and results can take a very long time. The forte of the NIRS technique lies in its speed [3]. *Author for correspondence NIRS is an analytical technique that uses no chemicals, gives accurate and precise results in minutes or even continuously, and is simple to install, and is safe to use. It is a rapid method for qualitative and quantitative analysis of a very wide range of materials, powders, slurries, and even solid materials and gases. A "rapid" method is defined as a method that will provide accurate and precise results in two minutes or less, including the time for sample preparation, and accessing the sample too, and removing it from the instrument. The NIRS spectrum analysis does not require samples to be grounded because when the light passes through it, the spectra of light can reveal a certain characteristics or features that are unique to the class of that sample.Artificial neural network (ANN) is a computational system that employs parallel processing based on connections and distributional methodology. ANN is similar in operations to a biological neural network [4,5]. It can be trained to learn and once trained it can be used as a very good classifier. Just like the

Watermelon ripeness detector using near infrared spectroscopy

“…It shows that among nondestructive techniques in classifying ripe and unripe watermelon, this study has the highest percentage in classification accuracy of 92.5 %. This result means that NIRS frequencies of 730, 760, 810, and 860 nm are a promising technology for watermelon ripeness classification, as in [15]- [17], [19]- [22], [27] for other non-destructive detectors. This paper's main contribution is the proof that a low-cost NIRS IC using its four frequencies can be used for determining the ripeness of watermelon.…”

“…Ripeness level of watermelon by its seed color: a) unripe watermelon, and b) ripe watermelon significant advantage of NIRS is that it is typically not necessary to prepare a sample, so the analysis is straightforward and very quick (between 15 and 90 seconds) [18]. The NIRS spectrum assessment does not allow samples' grounding because the spectra of light show some different properties in another sample when the light passes through it [19]. Infrared spectrometry is a technique commonly used by researchers to examine food and agricultural products [20].…”

Discrimination of civet coffee using visible spectroscopy

“…It can be seen in Table 6 Figure 2. Absorbance values of civet and non-civet coffee bean Analysis and Logistic Regression accurately classify the samples according to their group with 100 % accuracy, which is slightly better than [12] with 95 % to 100 % accuracy that uses ANN classification. This study successfully implements spectroscopy for civet coffee discrimination that used whole coffee bean and did not use chemicals in contrast with [4]- [11].…”

“…Arboleda [12] used Near-Infrared spectroscopy (NIRS), including Artificial Neural Network (ANN), to classify between civet coffee and Arabica coffee, which produced a classification result of 95% to 100% range. The study does not require grounded coffee beans as well as chemical mixtures.…”

Green coffee beans feature extractor using image processing

“…A notable exception is the use of near-infrared spectroscopy which does not need to destroy samples, requires very simple sample preparation, and has short span of time for feature extraction. Nevertheless, NIRS machines are expensive [25].…”