Pictorial depth cues are the visual information gathered from three dimensional scenes, and used to recover the third dimension of depth from two dimensional retinal images. Pictorial depth cues are also used to create the illusion of depth on pictorial representations, which are common platforms for architects to represent and visually examine the spatial qualities of their designs. Therefore, knowledge of pictorial depth cues can be used as a design strategy to imagine, depict, and enrich the spatial experience in architectural spaces.
The effect of pictorial depth cues is studied through psychophysical experiments. Measurements of participants' perceived distances can reveal the effect of depth cues in controlled experimental scenes, where depth information can be systematically varied. However, perceptual studies of depth cues are challenged by the dynamic character of the luminous environment in physical experimental settings. Therefore, the impact of luminance distribution patterns on depth perception is yet to be fully understood. In addition, restrictions of the displayable luminance range of common planar media hamper the realism offered by pictorial representations, and limit the study and applications of depth cues resulting from luminance distributions in architectural designs.
This dissertation draws from recent developments in computer graphics (physically based renderings and perceptually based tone-mapping techniques) and proposes a computational framework to generate pictorial spaces that can mimic the perceptual reality of architectural spaces. Psychophysical studies are conducted utilizing computer-generated images with the intent of establishing a cause-and-effect relationship between luminance distribution patterns in architectural configurations and the resultant perception of depth. The results of the studies demonstrate that luminance contrast is an effective depth cue that can either increase or decrease the perceived distances. Application of this pictorial depth cue in architectural design is demonstrated through the simulation and visualization of various architectural scenes.