Latent Image and Video Resolution Prediction using Convolutional Neural Networks

TL;DR

This paper addresses the latent resolution prediction problem in videos, proposing CNN-based methods that accurately estimate the original resolution of upscaled videos, which is crucial for quality assessment.

Contribution

It formulates the latent resolution prediction problem, creates a dataset, and introduces CNN algorithms that achieve high accuracy in predicting native video resolution.

Findings

CNN methods predict latent resolution with about 95% accuracy

The problem is formulated and dataset constructed for training and evaluation

Proposes new machine learning algorithms for latent resolution prediction

Abstract

This paper introduces a Video Quality Assessment (VQA) problem that has received little attention in the literature, called the latent resolution prediction problem. The problem arises when images or videos are upscaled from their native resolution and are reported as having a higher resolution than their native resolution. This paper formulates the problem, constructs a dataset for training and evaluation, and introduces several machine learning algorithms, including two Convolutional Neural Networks (CNNs), to address this problem. Experiments indicate that some proposed methods can predict the latent video resolution with about 95% accuracy.