Vlad Hosu

Shaolin Su, Hanhe Lin, Vlad Hosu et al.

An accurate computational model for image quality assessment (IQA) benefits many vision applications, such as image filtering, image processing, and image generation. Although the study of face images is an important subfield in computer vision research, the lack of face IQA data and models limits the precision of current IQA metrics on face image processing tasks such as face superresolution, face enhancement, and face editing. To narrow this gap, in this paper, we first introduce the largest annotated IQA database developed to date, which contains 20,000 human faces -- an order of magnitude larger than all existing rated datasets of faces -- of diverse individuals in highly varied circumstances. Based on the database, we further propose a novel deep learning model to accurately predict face image quality, which, for the first time, explores the use of generative priors for IQA. By taking advantage of rich statistics encoded in well pretrained off-the-shelf generative models, we obtain generative prior information and use it as latent references to facilitate blind IQA. The experimental results demonstrate both the value of the proposed dataset for face IQA and the superior performance of the proposed model.

Vlad Hosu, Marcos V. Conde, Lorenzo Agnolucci et al.

We introduce the AIM 2024 UHD-IQA Challenge, a competition to advance the No-Reference Image Quality Assessment (NR-IQA) task for modern, high-resolution photos. The challenge is based on the recently released UHD-IQA Benchmark Database, which comprises 6,073 UHD-1 (4K) images annotated with perceptual quality ratings from expert raters. Unlike previous NR-IQA datasets, UHD-IQA focuses on highly aesthetic photos of superior technical quality, reflecting the ever-increasing standards of digital photography. This challenge aims to develop efficient and effective NR-IQA models. Participants are tasked with creating novel architectures and training strategies to achieve high predictive performance on UHD-1 images within a computational budget of 50G MACs. This enables model deployment on edge devices and scalable processing of extensive image collections. Winners are determined based on a combination of performance metrics, including correlation measures (SRCC, PLCC, KRCC), absolute error metrics (MAE, RMSE), and computational efficiency (G MACs). To excel in this challenge, participants leverage techniques like knowledge distillation, low-precision inference, and multi-scale training. By pushing the boundaries of NR-IQA for high-resolution photos, the UHD-IQA Challenge aims to stimulate the development of practical models that can keep pace with the rapidly evolving landscape of digital photography. The innovative solutions emerging from this competition will have implications for various applications, from photo curation and enhancement to image compression.

Oliver Wiedemann, Vlad Hosu, Shaolin Su et al.

Scale-invariance is an open problem in many computer vision subfields. For example, object labels should remain constant across scales, yet model predictions diverge in many cases. This problem gets harder for tasks where the ground-truth labels change with the presentation scale. In image quality assessment (IQA), downsampling attenuates impairments, e.g., blurs or compression artifacts, which can positively affect the impression evoked in subjective studies. To accurately predict perceptual image quality, cross-resolution IQA methods must therefore account for resolution-dependent errors induced by model inadequacies as well as for the perceptual label shifts in the ground truth. We present the first study of its kind that disentangles and examines the two issues separately via KonX, a novel, carefully crafted cross-resolution IQA database. This paper contributes the following: 1. Through KonX, we provide empirical evidence of label shifts caused by changes in the presentation resolution. 2. We show that objective IQA methods have a scale bias, which reduces their predictive performance. 3. We propose a multi-scale and multi-column DNN architecture that improves performance over previous state-of-the-art IQA models for this task, including recent transformers. We thus both raise and address a novel research problem in image quality assessment.

Mariia Zameshina, Olivier Teytaud, Fabien Teytaud et al.

We design general-purpose algorithms for addressing fairness issues and mode collapse in generative modeling. More precisely, to design fair algorithms for as many sensitive variables as possible, including variables we might not be aware of, we assume no prior knowledge of sensitive variables: our algorithms use unsupervised fairness only, meaning no information related to the sensitive variables is used for our fairness-improving methods. All images of faces (even generated ones) have been removed to mitigate legal risks.

Vlad Hosu, Lorenzo Agnolucci, Daisuke Iso et al.

Image Quality Assessment (IQA) measures and predicts perceived image quality by human observers. Although recent studies have highlighted the critical influence that variations in the scale of an image have on its perceived quality, this relationship has not been systematically quantified. To bridge this gap, we introduce the Image Intrinsic Scale (IIS), defined as the largest scale where an image exhibits its highest perceived quality. We also present the Image Intrinsic Scale Assessment (IISA) task, which involves subjectively measuring and predicting the IIS based on human judgments. We develop a subjective annotation methodology and create the IISA-DB dataset, comprising 785 image-IIS pairs annotated by experts in a rigorously controlled crowdsourcing study. Furthermore, we propose WIISA (Weak-labeling for Image Intrinsic Scale Assessment), a strategy that leverages how the IIS of an image varies with downscaling to generate weak labels. Experiments show that applying WIISA during the training of several IQA methods adapted for IISA consistently improves the performance compared to using only ground-truth labels. The code, dataset, and pre-trained models are available at https://github.com/SonyResearch/IISA.

Feiran Li, Jiacheng Li, Marcos V. Conde et al.

We introduce the AIM 2025 Real-World RAW Image Denoising Challenge, aiming to advance efficient and effective denoising techniques grounded in data synthesis. The competition is built upon a newly established evaluation benchmark featuring challenging low-light noisy images captured in the wild using five different DSLR cameras. Participants are tasked with developing novel noise synthesis pipelines, network architectures, and training methodologies to achieve high performance across different camera models. Winners are determined based on a combination of performance metrics, including full-reference measures (PSNR, SSIM, LPIPS), and non-reference ones (ARNIQA, TOPIQ). By pushing the boundaries of camera-agnostic low-light RAW image denoising trained on synthetic data, the competition promotes the development of robust and practical models aligned with the rapid progress in digital photography. We expect the competition outcomes to influence multiple domains, from image restoration to night-time autonomous driving.

Vlad Hosu, Lorenzo Agnolucci, Oliver Wiedemann et al.

We introduce a novel Image Quality Assessment (IQA) dataset comprising 6073 UHD-1 (4K) images, annotated at a fixed width of 3840 pixels. Contrary to existing No-Reference (NR) IQA datasets, ours focuses on highly aesthetic photos of high technical quality, filling a gap in the literature. The images, carefully curated to exclude synthetic content, are sufficiently diverse to train general NR-IQA models. Importantly, the dataset is annotated with perceptual quality ratings obtained through a crowdsourcing study. Ten expert raters, comprising photographers and graphics artists, assessed each image at least twice in multiple sessions spanning several days, resulting in 20 highly reliable ratings per image. Annotators were rigorously selected based on several metrics, including self-consistency, to ensure their reliability. The dataset includes rich metadata with user and machine-generated tags from over 5,000 categories and popularity indicators such as favorites, likes, downloads, and views. With its unique characteristics, such as its focus on high-quality images, reliable crowdsourced annotations, and high annotation resolution, our dataset opens up new opportunities for advancing perceptual image quality assessment research and developing practical NR-IQA models that apply to modern photos. Our dataset is available at https://database.mmsp-kn.de/uhd-iqa-benchmark-database.html

Baptiste Roziere, Fabien Teytaud, Vlad Hosu et al.

We propose to use a quality estimator and evolutionary methods to search the latent space of generative adversarial networks trained on small, difficult datasets, or both. The new method leads to the generation of significantly higher quality images while preserving the original generator's diversity. Human raters preferred an image from the new version with frequency 83.7pc for Cats, 74pc for FashionGen, 70.4pc for Horses, and 69.2pc for Artworks, and minor improvements for the already excellent GANs for faces. This approach applies to any quality scorer and GAN generator.

Baptiste Roziere, Nathanal Carraz Rakotonirina, Vlad Hosu et al.

Super-resolution aims at increasing the resolution and level of detail within an image. The current state of the art in general single-image super-resolution is held by NESRGAN+, which injects a Gaussian noise after each residual layer at training time. In this paper, we harness evolutionary methods to improve NESRGAN+ by optimizing the noise injection at inference time. More precisely, we use Diagonal CMA to optimize the injected noise according to a novel criterion combining quality assessment and realism. Our results are validated by the PIRM perceptual score and a human study. Our method outperforms NESRGAN+ on several standard super-resolution datasets. More generally, our approach can be used to optimize any method based on noise injection.

Franz Götz-Hahn, Vlad Hosu, Dietmar Saupe

Data used to train supervised machine learning models are commonly split into independent training, validation, and test sets. This paper illustrates that complex data leakage cases have occurred in the no-reference image and video quality assessment literature. Recently, papers in several journals reported performance results well above the best in the field. However, our analysis shows that information from the test set was inappropriately used in the training process in different ways and that the claimed performance results cannot be achieved. When correcting for the data leakage, the performances of the approaches drop even below the state-of-the-art by a large margin. Additionally, we investigate end-to-end variations to the discussed approaches, which do not improve upon the original.

Franz Götz-Hahn, Vlad Hosu, Dietmar Saupe

In Neural Processing Letters 50,3 (2019) a machine learning approach to blind video quality assessment was proposed. It is based on temporal pooling of features of video frames, taken from the last pooling layer of deep convolutional neural networks. The method was validated on two established benchmark datasets and gave results far better than the previous state-of-the-art. In this letter we report the results from our careful reimplementations. The performance results, claimed in the paper, cannot be reached, and are even below the state-of-the-art by a large margin. We show that the originally reported wrong performance results are a consequence of two cases of data leakage. Information from outside the training dataset was used in the fine-tuning stage and in the model evaluation.

Hanhe Lin, Vlad Hosu, Dietmar Saupe

Multi-level deep-features have been driving state-of-the-art methods for aesthetics and image quality assessment (IQA). However, most IQA benchmarks are comprised of artificially distorted images, for which features derived from ImageNet under-perform. We propose a new IQA dataset and a weakly supervised feature learning approach to train features more suitable for IQA of artificially distorted images. The dataset, KADIS-700k, is far more extensive than similar works, consisting of 140,000 pristine images, 25 distortions types, totaling 700k distorted versions. Our weakly supervised feature learning is designed as a multi-task learning type training, using eleven existing full-reference IQA metrics as proxies for differential mean opinion scores. We also introduce a benchmark database, KADID-10k, of artificially degraded images, each subjectively annotated by 30 crowd workers. We make use of our derived image feature vectors for (no-reference) image quality assessment by training and testing a shallow regression network on this database and five other benchmark IQA databases. Our method, termed DeepFL-IQA, performs better than other feature-based no-reference IQA methods and also better than all tested full-reference IQA methods on KADID-10k. For the other five benchmark IQA databases, DeepFL-IQA matches the performance of the best existing end-to-end deep learning-based methods on average.

Hui Men, Vlad Hosu, Hanhe Lin et al.

Current benchmarks for optical flow algorithms evaluate the estimation either directly by comparing the predicted flow fields with the ground truth or indirectly by using the predicted flow fields for frame interpolation and then comparing the interpolated frames with the actual frames. In the latter case, objective quality measures such as the mean squared error are typically employed. However, it is well known that for image quality assessment, the actual quality experienced by the user cannot be fully deduced from such simple measures. Hence, we conducted a subjective quality assessment crowdscouring study for the interpolated frames provided by one of the optical flow benchmarks, the Middlebury benchmark. We collected forced-choice paired comparisons between interpolated images and corresponding ground truth. To increase the sensitivity of observers when judging minute difference in paired comparisons we introduced a new method to the field of full-reference quality assessment, called artefact amplification. From the crowdsourcing data, we reconstructed absolute quality scale values according to Thurstone's model. As a result, we obtained a re-ranking of the 155 participating algorithms w.r.t. the visual quality of the interpolated frames. This re-ranking not only shows the necessity of visual quality assessment as another evaluation metric for optical flow and frame interpolation benchmarks, the results also provide the ground truth for designing novel image quality assessment (IQA) methods dedicated to perceptual quality of interpolated images. As a first step, we proposed such a new full-reference method, called WAE-IQA. By weighing the local differences between an interpolated image and its ground truth WAE-IQA performed slightly better than the currently best FR-IQA approach from the literature.

Hanhe Lin, Vlad Hosu, Chunling Fan et al.

The satisfied user ratio (SUR) curve for a lossy image compression scheme, e.g., JPEG, characterizes the complementary cumulative distribution function of the just noticeable difference (JND), the smallest distortion level that can be perceived by a subject when a reference image is compared to a distorted one. A sequence of JNDs can be defined with a suitable successive choice of reference images. We propose the first deep learning approach to predict SUR curves. We show how to apply maximum likelihood estimation and the Anderson-Darling test to select a suitable parametric model for the distribution function. We then use deep feature learning to predict samples of the SUR curve and apply the method of least squares to fit the parametric model to the predicted samples. Our deep learning approach relies on a siamese convolutional neural network, transfer learning, and deep feature learning, using pairs consisting of a reference image and a compressed image for training. Experiments on the MCL-JCI dataset showed state-of-the-art performance. For example, the mean Bhattacharyya distances between the predicted and ground truth first, second, and third JND distributions were 0.0810, 0.0702, and 0.0522, respectively, and the corresponding average absolute differences of the peak signal-to-noise ratio at a median of the first JND distribution were 0.58, 0.69, and 0.58 dB. Further experiments on the JND-Pano dataset showed that the method transfers well to high resolution panoramic images viewed on head-mounted displays.

Franz Götz-Hahn, Vlad Hosu, Hanhe Lin et al.

Video quality assessment (VQA) methods focus on particular degradation types, usually artificially induced on a small set of reference videos. Hence, most traditional VQA methods under-perform in-the-wild. Deep learning approaches have had limited success due to the small size and diversity of existing VQA datasets, either artificial or authentically distorted. We introduce a new in-the-wild VQA dataset that is substantially larger and diverse: KonVid-150k. It consists of a coarsely annotated set of 153,841 videos having five quality ratings each, and 1,596 videos with a minimum of 89 ratings each. Additionally, we propose new efficient VQA approaches (MLSP-VQA) relying on multi-level spatially pooled deep-features (MLSP). They are exceptionally well suited for training at scale, compared to deep transfer learning approaches. Our best method, MLSP-VQA-FF, improves the Spearman rank-order correlation coefficient (SRCC) performance metric on the commonly used KoNViD-1k in-the-wild benchmark dataset to 0.82. It surpasses the best existing deep-learning model (0.80 SRCC) and hand-crafted feature-based method (0.78 SRCC). We further investigate how alternative approaches perform under different levels of label noise, and dataset size, showing that MLSP-VQA-FF is the overall best method for videos in-the-wild. Finally, we show that the MLSP-VQA models trained on KonVid-150k sets the new state-of-the-art for cross-test performance on KoNViD-1k, LIVE-VQC, and LIVE-Qualcomm with a 0.83, 0.75, and 0.64 SRCC, respectively. For both KoNViD-1k and LIVE-VQC this inter-dataset testing outperforms intra-dataset experiments, showing excellent generalization.

Vlad Hosu, Hanhe Lin, Tamas Sziranyi et al.

Deep learning methods for image quality assessment (IQA) are limited due to the small size of existing datasets. Extensive datasets require substantial resources both for generating publishable content and annotating it accurately. We present a systematic and scalable approach to creating KonIQ-10k, the largest IQA dataset to date, consisting of 10,073 quality scored images. It is the first in-the-wild database aiming for ecological validity, concerning the authenticity of distortions, the diversity of content, and quality-related indicators. Through the use of crowdsourcing, we obtained 1.2 million reliable quality ratings from 1,459 crowd workers, paving the way for more general IQA models. We propose a novel, deep learning model (KonCept512), to show an excellent generalization beyond the test set (0.921 SROCC), to the current state-of-the-art database LIVE-in-the-Wild (0.825 SROCC). The model derives its core performance from the InceptionResNet architecture, being trained at a higher resolution than previous models (512x384). Correlation analysis shows that KonCept512 performs similar to having 9 subjective scores for each test image.

Vlad Hosu, Bastian Goldlucke, Dietmar Saupe

We propose an effective deep learning approach to aesthetics quality assessment that relies on a new type of pre-trained features, and apply it to the AVA data set, the currently largest aesthetics database. While previous approaches miss some of the information in the original images, due to taking small crops, down-scaling or warping the originals during training, we propose the first method that efficiently supports full resolution images as an input, and can be trained on variable input sizes. This allows us to significantly improve upon the state of the art, increasing the Spearman rank-order correlation coefficient (SRCC) of ground-truth mean opinion scores (MOS) from the existing best reported of 0.612 to 0.756. To achieve this performance, we extract multi-level spatially pooled (MLSP) features from all convolutional blocks of a pre-trained InceptionResNet-v2 network, and train a custom shallow Convolutional Neural Network (CNN) architecture on these new features.

Hui Men, Hanhe Lin, Vlad Hosu et al.

Current benchmarks for optical flow algorithms evaluate the estimation quality by comparing their predicted flow field with the ground truth, and additionally may compare interpolated frames, based on these predictions, with the correct frames from the actual image sequences. For the latter comparisons, objective measures such as mean square errors are applied. However, for applications like image interpolation, the expected user's quality of experience cannot be fully deduced from such simple quality measures. Therefore, we conducted a subjective quality assessment study by crowdsourcing for the interpolated images provided in one of the optical flow benchmarks, the Middlebury benchmark. We used paired comparisons with forced choice and reconstructed absolute quality scale values according to Thurstone's model using the classical least squares method. The results give rise to a re-ranking of 141 participating algorithms w.r.t. visual quality of interpolated frames mostly based on optical flow estimation. Our re-ranking result shows the necessity of visual quality assessment as another evaluation metric for optical flow and frame interpolation benchmarks.

Hanhe Lin, Vlad Hosu, Dietmar Saupe

The main challenge in applying state-of-the-art deep learning methods to predict image quality in-the-wild is the relatively small size of existing quality scored datasets. The reason for the lack of larger datasets is the massive resources required in generating diverse and publishable content. We present a new systematic and scalable approach to create large-scale, authentic and diverse image datasets for Image Quality Assessment (IQA). We show how we built an IQA database, KonIQ-10k, consisting of 10,073 images, on which we performed very large scale crowdsourcing experiments in order to obtain reliable quality ratings from 1,467 crowd workers (1.2 million ratings). We argue for its ecological validity by analyzing the diversity of the dataset, by comparing it to state-of-the-art IQA databases, and by checking the reliability of our user studies.