Výsledky vyhledávání - "Tariq, Taimoor"

Report

Autor: Karpenko, Volodymyr, Tariq, Taimoor, Condor, Jorge, Didyk, Piotr

Modern deep-learning based super-resolution techniques process images and videos independently of the underlying content and viewing conditions. However, the sensitivity of the human visual system to image details changes depending on the underlying

Externí odkaz: http://arxiv.org/abs/2411.17513

Zobrazit plný text záznamu

Report

Noise-based Enhancement for Foveated Rendering

Autor: Tariq, Taimoor, Tursun, Cara, Didyk, Piotr

Human visual sensitivity to spatial details declines towards the periphery. Novel image synthesis techniques, so-called foveated rendering, exploit this observation and reduce the spatial resolution of synthesized images for the periphery, avoiding t

Externí odkaz: http://arxiv.org/abs/2204.04455

Zobrazit plný text záznamu

Report

A HVS-inspired Attention to Improve Loss Metrics for CNN-based Perception-Oriented Super-Resolution

Autor: Tariq, Taimoor, Gonzalez, Juan Luis, Kim, Munchurl

Deep Convolutional Neural Network (CNN) features have been demonstrated to be effective perceptual quality features. The perceptual loss, based on feature maps of pre-trained CNN's has proven to be remarkably effective for CNN based perceptual image

Externí odkaz: http://arxiv.org/abs/1904.00205

Zobrazit plný text záznamu

Report

Why Are Deep Representations Good Perceptual Quality Features?

Autor: Tariq, Taimoor, Tursun, Okan Tarhan, Kim, Munchurl, Didyk, Piotr

Recently, intermediate feature maps of pre-trained convolutional neural networks have shown significant perceptual quality improvements, when they are used in the loss function for training new networks. It is believed that these features are better

Externí odkaz: http://arxiv.org/abs/1812.00412

Zobrazit plný text záznamu