Color Texture Image Complexity—EEG-Sensed Human Brain Perception vs. Computed Measures

Abstract

In practical applications, such as patient brain signals monitoring, a non-invasive recording system with fewer channels for an easy setup and a wireless connection for remotely monitor physiological signals will be beneficial. In this paper, we investigate the feasibility of using such a system in a visual perception scenario. We investigate the complexity perception of color natural and synthetic fractal texture images, by studying the correlations between four types of data: image complexity that is expressed by computed color entropy and color fractal dimension, human subjective evaluation by scoring, and the measured brain EEG responses via Event-Related Potentials. We report on the considerable correlation experimentally observed between the recorded EEG signals and image complexity while considering three complexity levels, as well on the use of an EEG wireless system with few channels for practical applications, with the corresponding electrodes placement in accordance with the type of neural activity recorded.

On the Quantification of Visual Texture Complexity

“…The very slight changes in the correlation coefficients (r) in Figure 17 indicate that the number of frequencies (n F ) chosen to design the Gabor filter bank seems to have had little impact on image texture detection by the filter bank. A similar observation was made by Bianconi and Fernandez [34] for texture classification using Gabor filtering. They demonstrated that the number of frequencies and orientations used to design the filter bank did not significantly affect the accuracy of texture classification.…”

“…For the Gabor features, a bank of 36 Gabor filters were designed using six frequencies (n F = 6) (i.e., F ∈ {1, 0.7, 0.5, 0.35, 0.25, 0.18} with the maximum frequency F m = 1) and six orientation angles (n O = 6) i.e., θ ∈ {0, 30, 60, 90, 120, 150} • (see Figure 3). As recommended in [34], the frequency ratio F r was half-octave frequency spacing (F r = √ 2) and the smoothing parameters were γ = η = 0.5. The mean and standard deviation of the energy of the filtered images were averaged over the six orientation angles to obtain the rotation-invariant Gabor features of the textile images.…”

“…In one of their recent works, Nicolae and Ivanovici [34] investigated the complexity perception of natural texture images from the VisTex dataset [31] and synthetic fractal images. In addition to a subjective complexity scoring by eight young participants, electroencephalography (EEG) was conducted to record the participants' brain signals while observing the images.…”

How Does Image Complexity Affect the Accuracy of an Interactive Image Segmentation Algorithm?

“…Where pi is the probability of appearance of pixel value i in the image and N the amount of possible pixel values. Besides that, according to Nicolae and Ivanovici [13], many different entropy types and optimised versions have emerged in recent years, each with its own set of advantages, such as multi-scale entropy, cross entropy, fuzzy entropy, and many more when considering spatial information, with additional applications in the biomedical imaging domain.…”

“…According to Nicolae and Ivanovici [13], fractal dimension is the most representative measure for expressing the fractal geometry of colour texture images. The fractal dimension expresses texture variations and irregularities in relation to self-similar regions observed across different size scales.…”