Deep Portrait Quality Assessment. A NTIRE 2024 Challenge Survey

TL;DR

This paper surveys the NTIRE 2024 Portrait Quality Assessment Challenge, which evaluates deep neural networks' ability to assess portrait photo quality across diverse conditions, highlighting top solutions and current state-of-the-art performance.

Contribution

It provides a comprehensive review of the challenge, including proposed solutions, results, and insights into the effectiveness of current deep learning methods for portrait quality assessment.

Findings

Top models achieve high accuracy across diverse conditions

Deep neural networks can generalize well to challenging portrait scenarios

The challenge sets a new benchmark for portrait quality assessment

Abstract

This paper reviews the NTIRE 2024 Portrait Quality Assessment Challenge, highlighting the proposed solutions and results. This challenge aims to obtain an efficient deep neural network capable of estimating the perceptual quality of real portrait photos. The methods must generalize to diverse scenes and diverse lighting conditions (indoor, outdoor, low-light), movement, blur, and other challenging conditions. In the challenge, 140 participants registered, and 35 submitted results during the challenge period. The performance of the top 5 submissions is reviewed and provided here as a gauge for the current state-of-the-art in Portrait Quality Assessment.