In this article, we will discuss the image that whets the appetite of espresso consumers the most.
Tiger's Eye! Or Tiger's Back!
Is the tiger's back image, which is one of the expressions commonly used when describing a full espresso, and which even causes the espresso to look wrong if it is not there, an optical illusion? Or do our feelings towards espresso reflect the truth? Let's consider the answer to these questions from a scientific perspective.
This is not about espresso crema, as a very detailed article is about to be published for the analysis and evaluation of espresso crema. This is about the more specific topic of espresso surface appearance, the tiger image.
An important element in the evaluation of espresso is the crema, which is the golden brown foam layer formed on top of the espresso. Crema is an indicator that contributes to the espresso experience both visually and sensorily. An optimal extraction process is of critical importance to reveal the full flavor and characteristics of the coffee beans. This critical process provides negative or positive sensory contributions to the espresso, but also includes a subjective feeling in the evaluation of the resulting extract. When we see lemon, our mouth waters, when we hear vanilla, we perceive it as sweet, these are natural reflexes and judgments. However, these evaluation methods based on human judgment create inconsistencies and variability due to the nature of the process, which makes it difficult to obtain reliable and standard results. Therefore, the need for objective and reliable methods to provide more accurate and consistent evaluation of crema quality in coffee experiments is increasing. Because the more deeply science delves into a subject, the more neutral that subject becomes. Recently, machine learning techniques and computer vision technologies have come to the fore as promising solutions in various applications. Although machine learning techniques have great potential, widespread adoption of these technologies in the coffee industry is limited due to the lack of standard methods and integration problems with coffee machine systems. It is almost a benefactor that scientists also run machine learning for coffee science. In a study, analyses were performed on the cream images coming out of the machines with artificial intelligence-supported nanocomputer systems, thanks to extraction evaluation algorithms. In this study, unlike the existing methods that have been done before, an analysis method working with unsupervised learning was used to eliminate the lack of an Agtron-based analysis for real extraction results. Thus, by integrating this algorithm into the coffee machine hardware to provide an effective and fast evaluation, it was possible to make a real-time evaluation during the brewing process. The researchers went even further and considered the process from the coffee grinding process onwards and did not ignore the effect of particle distributions on espresso.
In the study, cream and extract were completely separated from each other, a series of analyses were performed by clustering the colors of the cream and optimizing these clusters. Color analyses were performed in a manner that everyone is familiar with. Of course, the Agtron Device. The K-Means clustering algorithm, one of the machine learning methods, was used to analyze the color of espresso cream. An RGB image consists of M x N pixels, each consisting of red, green and blue components. This algorithm provides an analysis in which each cluster contains similar colors by grouping pixels according to their color similarities. When evaluating the extraction results of espresso cream, especially the red (R) component and Agtron color values were compared. If we open a very small parenthesis here, the Agtron device, in principle, detects the surface colors of coffee beans using the light spectrum and expresses this color as an Agtron Value. It is one of the most important devices for coffee makers in coffee consistency, quality optimization and profiling. It makes businesses consistent by turning subjective approaches into objects. In the study, the color of the cream was analyzed with machine learning algorithms, taking into account different environmental conditions and cup types. After all, tiger tracks are formed due to a color difference, and being able to analyze the extracts on these tracks indicates whether the tiger tracks can really shine or not. Thanks to the Grabcut algorithm and the proposed polarization removal method, the color distribution of the cream was effectively separated and analyzed. In this process, high-resolution images optimized in square format and 500x500 pixels resolution were divided into color clusters by the K-Means algorithm. The RGB values obtained from the clustered colors helped to determine the dominant color tones of the cream and to understand whether the extraction was excessive or insufficient.
When we look at the results of the study, the clearest of the colorings used in the analysis were collected and presented visually. According to the result, the color of the espresso in the middle received a passing grade, while the ones on the right and left failed. According to this result, we have reached the conclusion that our tiger can also be comfortable in its place. As can be seen, if it is not exaggerated, that is, if it does not go into excessive extract, these stains on the espresso indicate that the espresso is at its most optimal level. It has been written in another article that the tiger's back image is the result of a channeling, which we are ignoring for now since it has not yet been based on a scientific infrastructure. When we look at this color perception with personal observation, the light reflected on the cream creates polarized effects, which even results in the cream not being perceived correctly. In fact, the stains on the espresso we see, which we take with phones or machines, are very different from the environment or the lights of the machines or the perceptions of our eyes. According to the study's evidence, tiger's back/eye are mottled, golden brown markings that occur when the espresso machine works under high pressure and temperature, releasing the coffee's oil and water-soluble components evenly, and are an indication that the espresso has been extracted very well. At the same time, the tiger's eye pattern means that the ideal balance between sweetness, bitterness and acidity has been achieved. Therefore, its appearance is not only an aesthetic element, but also an indicator of the taste profile. In fact, the researchers stated that these markings occur more in quality coffees.
Future directions include increasing algorithm efficiency, IoT integration for predictive maintenance and extension with sensor data, and developing sensory profiles with real-time cream evaluations. The most recent publication recorded a success rate of 97%. In these sensory advances, the use of artificial intelligence and sensory analysis seems likely to upset the royalists more than the king. Is it really necessary to define and name the senses with one-to-one facts? Does a sip of coffee taste different when it is drunk because it is enjoyable, and different when it is drunk by giving realistic names to its indicators such as taste, appearance, and texture?
DUYGU KURTULUS