หน้าหลัก

Esearch focused on low-level capabilities (as originally defined by Berman et al., 2014), exactly where, in an experimental setting, 52 study participants (26 female, imply age = 21.1) provided aesthetic preference ratings, making use of a 7-point rating scale for 307 images (naturalness ratings were similarly obtained from Berman et al., 2014). Participant aesthetic preference ratings have been then modeled making use of low-level visual attributes as predictors. It can be crucial to note that within the existing study, information from 260 of your original 307 images were utilised; 47 photos have been excluded because they have been vintage or presented in portrait (vs. landscape) orientation. The Hunter and Askarinejad (2015) analysis focused on taking a theoretically driven strategy to defining high-level semantic functions. They didn't, having said that, use these features to predict aesthetic preference and naturalness judgments. As such, these researchers deliver a helpful toolkit for operationalizing highlevel sematic functions, and in this study, we use those functions to predict aesthetic preference and naturalness judgments for the first time.Low-Level Visual FeaturesResearch from Berman et al. (2014) quantified 10 low-level visual attributes of environmental scene photos. These low-level attributes included spatial characteristics including edge density, straightedge density and entropy, and color characteristics for instance hue,Frontiers in Psychology | www.frontiersin.orgApril 2017 | Volume 8 | ArticleIbarra et al.Image Function Predictions of Preference and NaturalnessTABLE 1 | Colour and edge low-level characteristics quantified for aesthetic preference and naturalness models. Color Capabilities Hue (avg) Saturation (avg) Brightness (avg) SDhue (hue normal deviation) SDsaturation (saturation regular deviation) SDbrightness (brightness normal deviation) EDGE Capabilities Straight Edge Density Disorganized Edge Ratio Edge Density Entropysaturation, and brightness. Table 1 shows a total list of these low-level attributes. Berman et al. (2014) calculated the colour functions applying MATLAB's image processing toolbox's built-in functions (MATLAB and Image Processing Toolbox Release 2012b, The MathWorks, Inc., Natick, Massachusetts, United states). Hue (dominant wave length within the image), saturation (ratio of hue to other wavelengths in the image), and brightness (an image's color intensity--visibly, it is the amount of darkness/lightness in an image) have been calculated per pixel in every image, and those values were then averaged for each and every image to establish the image's average hue, saturation and brightness, respectively. According to the exact same pixel values, the standard deviations for every single color feature (SDhue, SDsaturation, SDbrightness) had been also calculated, which quantified the volume of diversity of these features in every image. For edge detection, Berman et al. (2014) utilized MATLAB's built-in "edge" function set to "canny." Canny edge detection makes use of a five-stage algorithm (Canny, 1986) to filter noise and track strong and weak edges. The edge density ratio was calculated as the ratio of edge pixels to total pixels for every single image. The pixels belonging to long straight edges were then distinguished from other edge pixels to quantify straight edge density, at the same time because the ratio with the non-straight edges to total edge content material which was labeled because the disorganized edge ratio of every single image (see Berman et al., 2014 for a lot more particulars). Lastly, gray-scale entropy was calculated from the histogram of gray-scale intensity values across 256.