No-reference video quality assessment (NR-VQA) for gaming videos is challenging due to limited human-rated datasets and unique content characteristics including fast motion, stylized graphics, and compression artifacts. We present MTL-VQA, a multi-task learning framework that uses full-reference metrics as supervisory signals to learn perceptually meaningful features without human labels for pretraining. By jointly optimizing multiple full-reference (FR) objectives with adaptive task weighting, our approach learns shared representations that transfer effectively to NR-VQA. Experiments on gaming video datasets show MTL-VQA achieves performance competitive with state-of-the-art NR-VQA methods across both MOS-supervised and label-efficient/self-supervised settings.

Photographs of people taken by professional photographers typically present the person in beautiful lighting, with an interesting pose, and flattering quality. This is unlike common photos people can take of themselves. In this paper, we explore how to create a “professional” version of a person's photograph, i.e., in a chosen pose, in a simple environment, with good lighting, and standard black top/bottom clothing. A key challenge is to preserve the person's unique identity, face and body features while transforming the photo. If there would exist a large paired dataset of the same person photographed both “in the wild” and by a professional photographer, the problem would potentially be easier to solve. However, such data does not exist, especially for a large variety of identities. To that end, we propose two key insights: 1) Our method transforms the input photo and person's face to a canonical UV space, which is further coupled with reposing methodology to model occlusions and novel view synthesis. Operating in UV space allows us to leverage existing unpaired datasets. 2) We personalize the output photo via multi image finetuning. Our approach yields high-quality, reposed portraits and achieves strong qualitative and quantitative performance on real-world imagery.

Evaluating the perceptual quality of Novel View Synthesis (NVS) images remains a key challenge, particularly in the absence of pixel-aligned ground truth references. Full-Reference Image Quality Assessment (FR-IQA) methods fail under misalignment, while No-Reference (NR-IQA) methods struggle with generalization. In this work, we introduce a Non-Aligned Reference (NAR-IQA) framework tailored for NVS, where it is assumed that the reference view shares partial scene content but lacks pixel-level alignment. We constructed a large-scale image dataset containing synthetic distortions targeting Temporal Regions of Interest (TROI) to train our NAR-IQA model. Our model is built on a contrastive learning framework that incorporates LoRA-enhanced DINOv2 embeddings and is guided by supervision from existing IQA methods. We train exclusively on synthetically generated distortions, deliberately avoiding overfitting to specific real NVS samples and thereby enhancing the model's generalization capability. Our model outperforms state-of-the-art FR-IQA, NR-IQA, and NAR-IQA methods, achieving robust performance on both aligned and non-aligned references. We also conducted a novel user study to gather data on human preferences when viewing non-aligned references in NVS. We find strong correlation between our proposed quality prediction model and the collected subjective ratings. For dataset and code, please visit our project page: https://stootaghaj.github.io/nova-project/

Banding artifacts, or false contouring, are a common compression impairment that often appears on large smooth regions of encoded videos and images. These staircase-like color bands can be very noticeable and annoying, even on otherwise high-quality videos, especially when displayed on high-definition screens. Yet, relatively little attention has been applied to this problem. Here we study this artifact, by first analyzing the perceptual and encoding aspects of banding artifacts, then propose a new distortion-specific no-reference video quality algorithm for predicting banding artifacts, inspired by perceptual models. The proposed banding detector can generate a pixel-wise banding visibility map, and output overall banding severity scores at both the frame and video levels. Furthermore, we propose a deep learning based approach to improve the overall perceptual quality of compressed videos by joint debanding and compression artifact removal. Our experimental results show that the proposed banding detector delivers better consistency with subjective evaluations, and is able to detect different perceptual severity levels of bands. The debanding experiments also show that the proposed algorithm outperforms recent debanding models both visually and quantitatively. The code is available at https://github.com/google/bband-adaband and https://github.com/vztu/DebandingNet.

Image Quality Assessment (IQA) models aim to predict perceptual image quality in alignment with human judgments. No-Reference (NR) IQA remains particularly challenging due to the absence of a reference image. While deep learning has significantly advanced this field, a major hurdle in developing NR-IQA models is the limited availability of subjectively labeled data. Most existing deep learning-based NR-IQA approaches rely on pre-training on large-scale datasets before fine-tuning for IQA tasks. To further advance progress in this area, we propose a novel approach that constructs a custom dataset using a limited number of reference content images and introduces a no-reference IQA model that incorporates both content and quality features for perceptual quality prediction. Specifically, we train a quality-aware model using contrastive triplet-based learning, enabling efficient training with fewer samples while achieving strong generalization performance across publicly available datasets. Our repository is available at https://github.com/rajeshsureddi/triqa.

Implicit Neural Representations (INRs) have been demonstrated to achieve state-of-the-art compression of a broad range of modalities such as images, videos, 3D surfaces, and audio. Most studies have focused on building neural counterparts of traditional implicit representations of 3D geometries, such as signed distance functions. However, the triangle mesh-based representation of geometry remains the most widely used representation in the industry, while building INRs capable of generating them has been sparsely studied. In this paper, we present a method for building compact INRs of zero-genus 3D manifolds. Our method relies on creating a spherical parameterization of a given 3D mesh - mapping the surface of a mesh to that of a unit sphere - then constructing an INR that encodes the displacement vector field defined continuously on its surface that regenerates the original shape. The compactness of our representation can be attributed to its hierarchical structure, wherein it first recovers the coarse structure of the encoded surface before adding high-frequency details to it. Once the INR is computed, 3D meshes of arbitrary resolution/connectivity can be decoded from it. The decoding can be performed in real time while achieving a state-of-the-art trade-off between reconstruction quality and the size of the compressed representations.

This paper reports IEEE International Conference on Multimedia & Expo (ICME) 2025 Grand Challenge on Generalizable HDR and SDR Video Quality Measurement. With the rapid development of video technology, especially High Dynamic Range (HDR) and Standard Dynamic Range (SDR) contents, the need for robust and generalizable Video Quality Assessment (VQA) methods has become increasingly demanded. Existing VQA models often struggle to deliver consistent performance across varying dynamic ranges, distortion types, and diverse content. This challenge was established to benchmark and promote VQA approaches capable of jointly handling HDR and SDR content. In the final evaluation phase, five teams submitted seven models along with technical reports to the Full Reference (FR) and No Reference (NR) tracks. Among them, four methods outperformed VMAF baseline, while the top-performing model achieved state-of-the-art performance, setting a new benchmark for generalizable video quality assessment.

Video service providers need their delivery systems to be able to adapt to network conditions, user preferences, display settings, and other factors. HTTP Adaptive Streaming (HAS) offers dynamic switching between different video representations to simultaneously enhance bandwidth consumption and users' streaming experiences. Per-shot encoding, pioneered by Netflix, optimizes the encoding parameters on each scene or shot. The Dynamic Optimizer (DO) uses the Video Multi-Method Assessment Fusion (VMAF) perceptual video quality prediction engine to deliver high-quality videos at reduced bitrates. Here we develop a perceptually optimized method of constructing optimal per-shot bitrate and quality ladders, using an ensemble of low-level features and Visual Information Fidelity (VIF) features. During inference, our method predicts the bitrate or quality ladder of a source video without any compression or quality estimation. We compare the performance of our model against other content-adaptive bitrate ladder prediction methods, a fixed bitrate ladder, and reference bitrate ladders constructed via exhaustive encoding using Bjøntegaard-delta (BD) metrics. Our proposed method shows excellent gains in bitrate and quality against the fixed bitrate ladder and only small losses against the reference bitrate ladder, while providing significant computational advantages.

Image Quality Assessment (IQA) models are employed in many practical image and video processing pipelines to reduce storage, minimize transmission costs, and improve the Quality of Experience (QoE) of millions of viewers. These models are sensitive to a diverse range of image distortions and can accurately predict image quality as judged by human viewers. Recent advancements in generative models have resulted in a significant influx of "GenAI" content on the internet. Existing methods for detecting GenAI content have progressed significantly with improved generalization performance on images from unseen generative models. Here, we leverage the capabilities of existing IQA models, which effectively capture the manifold of real images within a bandpass statistical space, to distinguish between real and AI-generated images. We investigate the generalization ability of these perceptual classifiers to the task of GenAI image detection and evaluate their robustness against various image degradations. Our results show that a two-layer network trained on the feature space of IQA models demonstrates state-of-the-art performance in detecting fake images across generative models, while maintaining significant robustness against image degradations.

We introduce HDRSDR-VQA, a large-scale video quality assessment dataset designed to facilitate comparative analysis between High Dynamic Range (HDR) and Standard Dynamic Range (SDR) content under realistic viewing conditions. The dataset comprises 960 videos generated from 54 diverse source sequences, each presented in both HDR and SDR formats across nine distortion levels. To obtain reliable perceptual quality scores, we conducted a comprehensive subjective study involving 145 participants and six consumer-grade HDR-capable televisions. A total of over 22,000 pairwise comparisons were collected and scaled into Just-Objectionable-Difference (JOD) scores. Unlike prior datasets that focus on a single dynamic range format or use limited evaluation protocols, HDRSDR-VQA enables direct content-level comparison between HDR and SDR versions, supporting detailed investigations into when and why one format is preferred over the other. The open-sourced part of the dataset is publicly available to support further research in video quality assessment, content-adaptive streaming, and perceptual model development.

Despite recent advancements in latent diffusion models that generate high-dimensional image data and perform various downstream tasks, there has been little exploration into perceptual consistency within these models on the task of No-Reference Image Quality Assessment (NR-IQA). In this paper, we hypothesize that latent diffusion models implicitly exhibit perceptually consistent local regions within the data manifold. We leverage this insight to guide on-manifold sampling using perceptual features and input measurements. Specifically, we propose Perceptual Manifold Guidance (PMG), an algorithm that utilizes pretrained latent diffusion models and perceptual quality features to obtain perceptually consistent multi-scale and multi-timestep feature maps from the denoising U-Net. We empirically demonstrate that these hyperfeatures exhibit high correlation with human perception in IQA tasks. Our method can be applied to any existing pretrained latent diffusion model and is straightforward to integrate. To the best of our knowledge, this paper is the first work on guiding diffusion model with perceptual features for NR-IQA. Extensive experiments on IQA datasets show that our method, LGDM, achieves state-of-the-art performance, underscoring the superior generalization capabilities of diffusion models for NR-IQA tasks.

Facial expressions are key to human communication, conveying emotions and intentions. Given the rising popularity of digital humans and avatars, the ability to accurately represent facial expressions in real time has become an important topic. However, quantifying perceived differences between pairs of expressions is difficult, and no comprehensive subjective datasets are available for testing. This work introduces a new dataset targeting this problem: FaceExpressions-70k. Obtained via crowdsourcing, our dataset contains 70,500 subjective expression comparisons rated by over 1,000 study participants1 We demonstrate the applicability of the dataset for training perceptual expression difference models and guiding decisions on acceptable latency and sampling rates for facial expressions when driving a face avatar.

High Dynamic Range (HDR) videos have enjoyed a surge in popularity in recent years due to their ability to represent a wider range of contrast and color than Standard Dynamic Range (SDR) videos. Although HDR video capture has seen increasing popularity because of recent flagship mobile phones such as Apple iPhones, Google Pixels, and Samsung Galaxy phones, a broad swath of consumers still utilize legacy SDR displays that are unable to display HDR videos. As result, HDR videos must be processed, i.e., tone-mapped, before streaming to a large section of SDR-capable video consumers. However, server-side tone-mapping involves automating decisions regarding the choices of tone-mapping operators (TMOs) and their parameters to yield high-fidelity outputs. Moreover, these choices must be balanced against the effects of lossy compression, which is ubiquitous in streaming scenarios. In this work, we develop a novel, efficient model of objective video quality named Cut-FUNQUE that is able to accurately predict the visual quality of tone-mapped and compressed HDR videos. Finally, we evaluate Cut-FUNQUE on a large-scale crowdsourced database of such videos and show that it achieves state-of-the-art accuracy.

We describe a search-free resizing framework that can further improve the rate-distortion tradeoff of recent learned image compression models. Our approach is simple: compose a pair of differentiable downsampling/upsampling layers that sandwich a neural compression model. To determine resize factors for different inputs, we utilize another neural network jointly trained with the compression model, with the end goal of minimizing the rate-distortion objective. Our results suggest that"compression friendly"downsampled representations can be quickly determined during encoding by using an auxiliary network and differentiable image warping. By conducting extensive experimental tests on existing deep image compression models, we show results that our new resizing parameter estimation framework can provide Bj{\o}ntegaard-Delta rate (BD-rate) improvement of about 10% against leading perceptual quality engines. We also carried out a subjective quality study, the results of which show that our new approach yields favorable compressed images. To facilitate reproducible research in this direction, the implementation used in this paper is being made freely available online at: https://github.com/treammm/ResizeCompression.

High-resolution texture maps are necessary for representing real-world objects accurately with 3D meshes. The large sizes of textures can bottleneck the real-time rendering of high-quality virtual 3D scenes on devices having low computational budgets and limited memory. Downsampling the texture maps directly addresses the issue, albeit at the cost of visual fidelity. Traditionally, downsampling of texture maps is performed using methods like bicubic interpolation and the Lanczos algorithm. These methods ignore the geometric layout of the mesh and its UV parametrization and also do not account for the rendering process used to obtain the final visualization that the users will experience. Towards filling these gaps, we introduce GeoScaler, which is a method of downsampling texture maps of 3D meshes while incorporating geometric cues, and by maximizing the visual fidelity of the rendered views of the textured meshes. We show that the textures generated by GeoScaler deliver significantly better quality rendered images compared to those generated by traditional downsampling methods

In recent years, deep learning-based methods have made significant progress on the image quality assessment problem; however, challenges remain arising from the lack of annotated, real-world training data and consequent poor generalization ability. Towards addressing these challenges, we propose a no-reference image quality assessment (NR-IQA) method based on generative AI (GenAI) images. Specifically, we use GenAI images as reference images, employing a cold diffusion model to generate distorted images of four different distortion types, and we label these distorted images using a full-reference model, thereby making it possible to construct a large-scale pre-training dataset. We use this resource generation method to facilitate NR-IQA model building. We deploy a Multi-scale Cross Attention Block (MCAB) and a Scale Simple Attention Module (SSAM) to enhance feature representation by extracting multi-scale feature information from both the channel and spatial dimensions that are predictive of image quality. Extensive experiments on eight public databases demonstrate that the proposed method achieves state-of-the-art (SOTA) performance. A public release of all the codes associated with this work will be made available on GitHub.

High Dynamic Range (HDR) videos enhance visual experiences with superior brightness, contrast, and color depth. The surge of User-Generated Content (UGC) on platforms like YouTube and TikTok introduces unique challenges for HDR video quality assessment (VQA) due to diverse capture conditions, editing artifacts, and compression distortions. Existing HDR-VQA datasets primarily focus on professionally generated content (PGC), leaving a gap in understanding real-world UGC-HDR degradations. To address this, we introduce CHUG: Crowdsourced User-Generated HDR Video Quality Dataset, the first large-scale subjective study on UGC-HDR quality. CHUG comprises 856 UGC-HDR source videos, transcoded across multiple resolutions and bitrates to simulate real-world scenarios, totaling 5,992 videos. A large-scale study via Amazon Mechanical Turk collected 211,848 perceptual ratings. CHUG provides a benchmark for analyzing UGC-specific distortions in HDR videos. We anticipate CHUG will advance No-Reference (NR) HDR-VQA research by offering a large-scale, diverse, and real-world UGC dataset. The dataset is publicly available at: https://shreshthsaini.github.io/CHUG/.

Implicit neural representations (INRs) have been successfully used to compress a variety of 3D surface representations such as Signed Distance Functions (SDFs), voxel grids, and also other forms of structured data such as images, videos, and audio. However, these methods have been limited in their application to unstructured data such as 3D meshes and point clouds. This work presents a simple yet effective method that extends the usage of INRs to compress 3D triangle meshes. Our method encodes a displacement field that refines the coarse version of the 3D mesh surface to be compressed using a small neural network. Once trained, the neural network weights occupy much lower memory than the displacement field or the original surface. We show that our method is capable of preserving intricate geometric textures and demonstrates state-of-the-art performance for compression ratios ranging from 4x to 380x (See Figure 1 for an example).

Motion-preserved video editing is crucial for creators, particularly in scenarios that demand flexibility in both the structure and semantics of swapped objects. Despite its potential, this area remains underexplored. Existing diffusion-based editing methods excel in structure-preserving tasks, using dense guidance signals to ensure content integrity. While some recent methods attempt to address structure-variable editing, they often suffer from issues such as temporal inconsistency, subject identity drift, and the need for human intervention. To address these challenges, we introduce VidMP3, a novel approach that leverages pose and position priors to learn a generalized motion representation from source videos. Our method enables the generation of new videos that maintain the original motion while allowing for structural and semantic flexibility. Both qualitative and quantitative evaluations demonstrate the superiority of our approach over existing methods.