Emotions, Learning, and Electrooculography
Another application in learning that an eye-tracking device based on electrooculography could have is in the development of a system for emotion recognition and assessment. It has been studied that emotions and learning have a strong relationship, according to King \cite{King_2019}, studies in disciplines such as neuroscience, physiology, and education have generated research that supports the association of emotions and learning. There have been several types of approaches to detect or classify emotions that go from electroencephalography, facial images processing, and even speech analysis through voice recordings, nonetheless, according to Lim \cite{Lim_2020}, the used of eye-tracking to classify or detect emotions is a field that is currently starting to be researched. In the case of EOG signal measurement, there has been some interest in the detection of emotions through this technique special since emotions play a crucial part in human communications \cite{Ekman_2005}. For instance, Paul \cite{Paul_2017}, elaborate a machine learning classification algorithm for emotion recognition based on EOG. It is concluded that EOG based emotion recognition techniques have some potential in generating real-time emotion evaluation system. Another similar work was presented by Soundariya\cite{Soundariya_2017} and \cite{Ma_2018} in which the relationship between electrooculography and emotions was also studied. Despite, there is a great number of studies that have tried to find a correlation between emotions and education or eye-tracking (EOG) and education, there is no research that has tried to merge education, electrooculography, and emotions. Therefore, this is an area of opportunity that must no be underestimated that could be implemented with a low-cost device as the one presented.
Advantages and disadvantages of the eye-tracking system.
It this section a brief comparison of the proposed eye-tracking system with the conventional ones and commercial is made to display its core characteristics and areas of possible improvement. The first disadvantage of the proposed eye tracking device is that it is more invasive than camera-based recognition systems. This is because the EOG technique requires the use of electrodes that must be placed on the user's skin, while the camera only requires that it be aimed at the user's eyes. In addition, the electrode placed on the skin must be replaced every time it is used, so a component of the system must be changed every time it is planned to use. Likewise, the manufacture of the electrode, design, and quality becomes important and deterministic points when using the electrooculography technique. Also, since the EOG is a physiological signal it is susceptible to artifacts, ambient, and body noise which sometimes makes the behavior of the EOG technique unpredictable in different environments or persons. Nonetheless, the majority of these problems can be avoided or corrected by applying the proposed signal processing methodology described in the previous section, where it is important to have a filtering stage, the feature extraction process, and a classification model.
On the other hand, because the processing method used to classify or monitor biopotentials turns out to be computationally less expensive than to process video or image, the proposed eye-tracking system can be implemented at a more accessible cost and without the need to require specialized components or parts that make it impossible to implement them independently by teachers or students. Therefore, it is possible to implement the processing techniques in other electronic embedded systems without the need for specialized hardware and software. This is supported by the software that was proposed in the previous sections of this in which Python was selected as the programming language that has an open license ar has a multiplatform development environment. Also, the microcontroller used for the sampling of the EOG signal is an Atmega328p with is the microcontroller in which the Arduino development platform is based. The Arduino platform, similar to the Raspberry has been widely used in educational institutions and research to motivate the learning of electronics and programming topics into the students \cite{Novak_2018}. By taking into account the above, it could feasible for students or professors to implement the eye-tracking device by only following the presented methodology and in this way enhance the development of other skills in the field of programming and electronics engineering. Also, EOG can be used as a substitute for Brain-Computer Interface since it can assess in some degree cognitive activities as established in the previous section of this work, but with a less quantity of electrodes, it can be adjusted by the user, and require less specialized hardware and software to make a the eye-tracking processing and classification.