Showing items related by title, author, creator and subject.

• 

  Bias effects of short- and long-term color memory for unique objects

  Bloj, Marina; Weiß, D.; Gegenfurtner, K.R. (09/03/2016)

  Are objects remembered with a more saturated color? Some of the evidence supporting this statement comes from research using “memory colors”—the typical colors of particular objects, for example, the green of grass. The problematic aspect of these findings is that many different exemplars exist, some of which might exhibit a higher saturation than the one measured by the experimenter. Here we avoid this problem by using unique personal items and comparing long- and short-term color memory matches (in hue, value, and chroma) with those obtained with the object present. Our results, on average, confirm that objects are remembered as more saturated than they are.
• 

  An electrophysiological study of chromatic processing in the human visual system. Using visual evoked potentials and electroretinograms to study cortical and retinal contributions to human trichromatic vision.

  McKeefry, Declan J.; Challa, Naveen K. (University of BradfordBradford School of Optometry and Vision Science, 2012-01-24)

  The work in this thesis is concerned with examining the retinal and cortical contributions to human trichromatic colour vision. Chromatic processing at the cortex level was examined using visual evoked potentials (VEPs). These responses were elicited by chromatic spot stimuli, which were manipulated in order to selectively activate the chromatic processing system. Chromatic processing at the retinal level was examined using the electroretinograms (ERGs) for which cone isolating stimuli were used to assess the nature of L and M cone inputs to cone-opponent mechanisms. The results from the VEP experiments suggest VEP morphology is dependent upon 1) chromatic and or luminance contrast content of the stimulus, 2) stimulus size, and 3) extent to which the chromatic stimulus activates either the L/M or S/(L+M) opponent mechanism. The experiments indicate that chromatic stimulation is indexed by large N1 component and small offset responses. Optimal stimulus size for chromatic isolation is 2-4 ° along L/M axes and 6° along S/(L+M) axis. From the ERG experiments, It has been shown that the low (12Hz) and high (30Hz) temporal frequency flickering stimuli can isolate the chromatic and luminance processing mechanisms in the retina. For low temporal frequency ERGs, the L:M ratio was close to unity and L/M phase difference was close to 180°. For high temporal frequency ERGs, the L:M ratio was more than unity and L/M phase difference was close to 90°. In addition to this, the variation in L:M ratio across the retinal eccentricity was also examined. These results suggest, for the chromatic processing, L:M ratio is close to unity independent of retinal eccentricity and individuals. For the luminance processing, L:M ratio is more than unity and depends upon the region of the retina being stimulated. These findings indicate the maintenance of cone selective input for the chromatic processing across the human retina.
• 

  Machine Learning for 3D Visualisation Using Generative Models

  Ugail, Hassan; Mehmoud, Irfan; Taif, Khasrouf M.M.; Yarmouk University (University of BradfordFaculty of Engineering and Informatics. School of Media, Design and Technology, 2020)

  One of the state-of-the-art highlights of deep learning in the past ten years is the introduction of generative adversarial networks (GANs), which had achieved great success in their ability to generate images comparable to real photos with minimum human intervention. These networks can generalise to a multitude of desired outputs, especially in image-to-image problems and image syntheses. This thesis proposes a computer graphics pipeline for 3D rendering by utilising generative adversarial networks (GANs). This thesis is motivated by regression models and convolutional neural networks (ConvNets) such as U-Net architectures, which can be directed to generate realistic global illumination effects, by using a semi-supervised GANs model (Pix2pix) that is comprised of PatchGAN and conditional GAN which is then accompanied by a U-Net structure. Pix2pix had been chosen for this thesis for its ability for training as well as the quality of the output images. It is also different from other forms of GANs by utilising colour labels, which enables further control and consistency of the geometries that comprises the output image. The series of experiments were carried out with laboratory created image sets, to pursue the possibility of which deep learning and generative adversarial networks can lend a hand to enhance the pipeline and speed up the 3D rendering process. First, ConvNet is applied in combination with Support Vector Machine (SVM) in order to pair 3D objects with their corresponding shadows, which can be applied in Augmenter Reality (AR) scenarios. Second, a GANs approach is presented to generate shadows for non-shadowed 3D models, which can also be beneficial in AR scenarios. Third, the possibility of generating high quality renders of image sequences from low polygon density 3D models using GANs. Finally, the possibility to enhance visual coherence of the output image sequences of GAN by utilising multi-colour labels. The results of the adopted GANs model were able to generate realistic outputs comparable to the lab generated 3D rendered ground-truth and control group output images with plausible scores on PSNR and SSIM similarity index metrices.