The Function of Color and Structure Based on EEG Features in Landscape Recognition

Abstract

Both color and structure make important contributions to human visual perception, as well as the evaluation of landscape quality and landscape aesthetics. The EEG equipment liveamp32 was used to record the EEG signals of humans when viewing landscape images, structure images with filtered color, and color images with a filtered structure. The results show that the SVM classifier was the most suitable classifier for landscape classification based on EEG features. The classification accuracy of the landscape picture recognition was up to 98.3% when using beta waves, while the accuracy of the color recognition was 97.5%, and that of the structure recognition was 93.9% when using gamma waves. Secondly, color and structure played a major role in determining the alpha and gamma wave responses, respectively, for all the landscape types, including forest, desert, and water. Furthermore, structure only played a decisive role in forest, while color played a major role in desert and water when using beta waves. Lastly, statistically significant differences between landscape groups and scenario groups with regard to alpha, beta, and gamma rhythms in brain waves were confirmed. The reasonable usage and layout of structure and color will have a very important guiding value for landscape aesthetics in future landscape design and landscape planning.

Landscape Perception Identification and Classification Based on Electroencephalogram (EEG) Features

“…At the same time, it also proves that it is feasible to identify people’s perception of different landscapes through multi-dimensional EEG features [ 60 ]. This result is of great significance for evaluating the impact of landscapes on people, which also has important reference value for the quantification of ecosystem cultural services [ 44 , 61 ].…”

Modeling the Visual Landscape: A Review on Approaches, Methods and Techniques

“…Information on brain visual selection based on combined data from fMRI, ERPs, and transcranial magnetic stimulation (TMS) is thoroughly discussed by Theeuwes [176]. More specifically, the brain activities of landscape perception and cognition have been recently explored by applying fMRI (e.g., [43][44][45][46][47]) and EEG techniques (e.g., [48][49][50][51][52][53]). It should be noticed that when utilizing these techniques it is more difficult to discern if the visual perception or cognition are modeled.…”

“…Regarding EEG, the research work of [49] aimed at assessing the accuracy of landscape perception and recognition of some typical landscape types (e.g., forest, wetland, and farmland) based on different EEG features and different classifiers. Furthermore, Wang et al [53] investigated the role of color and structure in landscape recognition by using the objective quantitative index of EEG features and ended up choosing a specific classifier (SVM) to properly study landscape recognition and aesthetics [53]. Roe et al [48] utilized objective indicators based on EEG data to detect emotional changes while viewing 'green' and 'gray' landscape scenes; these indicators were found to be consistent with subjective preferences, as well as with restorative theory.…”

“…Towards this direction, several research studies have recently been dedicated to the experimental recording and quantification of landscape visual perception, using mainly methods and sensor-based techniques known as eye tracking and eye movement analysis (e.g., [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42]). Other research studies have employed other sensor-based experimental techniques and technologies to register and quantify the brain activity during the observation of landscapes (scenes) such as functional Magnetic Resonance Imaging (fMRI) (e.g., [43][44][45][46][47]) and Electroencephalography (EEG) (e.g., [48][49][50][51][52][53]).…”