
Road markings have been standardized for human perception for over a century. With the rapid expansion of autonomous vehicles that rely on machine vision, new challenges emerge: markings optimized for human drivers may fail automated perception systems under low lighting, adverse weather, or high retroreflectivity conditions. Drawing on imaging science, vehicle dynamics, and experimental results from road-paint sample analysis, we identify four critical design factors, including spatial characteristics, color, contrast, and retroreflectivity, and provide concrete recommendations for evolving road marking standards. These augment rather than replace existing human-centric requirements and are developed in alignment with the IEEE P2020 Automotive Image Quality Standards working group.

This paper presents brilliantISP, a modular, open-source HDR image signal processing pipeline for automotive camera applications. Unlike existing open-source ISPs, which employ floating-point arithmetic and are not optimized for HDR automotive use cases, brilliantISP adopts a predominantly fixed-point, unsigned integer architecture with explicit, bounded bit depths at each processing stage, mirroring the constraints of production embedded ISPs while remaining accessible for research and experimentation. The pipeline incorporates a configurable decompanding stage that reconstructs a linear-domain signal from piecewise-companded sensor outputs, supporting sensors with effective dynamic ranges up to 144 dB. Multiple global tone mapping operators are provided, including Reinhard, ACES, and Hable, alongside a Durand-style local tone mapping operator that decomposes the image into base and detail layers for contrast-preserving dynamic range compression. Additional pipeline stages include defect pixel correction, black level correction, lens shading correction, auto white balance, a choice of six demosaicing algorithms, local contrast and edge enhancement, and gamma correction. All stages are configurable via YAML parameter files, and comprehensive debug logging provides block-level execution statistics, dynamic range metrics, bit depth utilization, and histogram outputs to support both algorithm development and ISP tuning studies. The pipeline is validated on imagery from a Sony IMX623 split-pixel HDR fisheye sensor, where decompanded input spans approximately 19.26 EV at 20.7-bit effective depth, compressed to a 3.01 EV, 7.9-bit output after tone mapping and gamma correction. BrilliantISP is intended as a practical research platform for studying HDR tone mapping, demosaicing, and ISP tuning in the context of automotive computational photography.

In order to determine the lowest light level at which a digital camera can still deliver acceptable images, acceptance thresholds must be established for all related image quality factors. ISO 19093 describes these factors and how they can be measured. However, the acceptance thresholds may depend on the application for which the images were captured, as well as on people's individual tolerance for degradation in the different image quality factors. To generate a standard set of tolerance levels for photographic applications, a psychophysical experiment was performed, as described in this paper. First, a group of 23 image quality experts participated, followed by 16 people with no specific experience in imaging. The same experiment was repeated for the specific application of security cameras. The set of images, as well as the questions asked of the participants, were adapted to the use case. For the security application, 27 participants with a background in security camera imaging took part.

Distortions introduced during the reproduction of digital images can lead to substantial changes in their color composition. The motivations for altering images range from practical purposes, such as image compression and color quantization to reduce file size, to more aesthetic applications like style transfer using generative AI. In this work, we investigate how the reproduction of color images affects material appearance, in particular, the perception of gloss and translucency. We applied different image quality distortions to natural images of glossy and translucent objects. Additionally, we Ghiblified them – a recent viral social media phenomenon of mimicking the Japanese anime style using generative AI style transfer. Afterward, we conducted a series of user studies to evaluate the fidelity of gloss and translucency reproduction. The experimental results represent how the reproductions are perceived by image quality metrics and open up a new direction for material appearance studies.

The National Gallery of Art developed a systematic approach to evaluate and categorize its extensive digital image collection spanning 20 years of technological evolution. This study addresses the challenge of inconsistent image quality resulting from varying capture technologies and methodologies over time. A four-tier rating system was created based on comprehensive analysis of capture devices, technical specifications, and workflow documentation. The system enables efficient assessment of image suitability for different applications while providing clear guidance for re-digitization decisions. The implementation includes integration with the institution's digital asset management system, offering a practical framework that other cultural heritage institutions can adapt for managing legacy digital collections while maintaining current quality standards.

This research investigated the influence of lightness, lightness contrast, observer characteristics, and display types on image preference and perception. Previous studies have emphasized the importance of color attributes in shaping image quality; in this study, we explored lightness attributes using CIECAM16 color space. Four experiments were conducted on OLED and QLED displays, during which participants adjusted color attributes of images to their preference and rated their preferred images relative to reference images. The results indicated that lightness attributes significantly impact image preference, and that observer characteristics and image content influence lightness preference on each display type.

Quantification of the image sensor signal and noise is essential to derive key image quality performance indicators. Image sensors in automotive cameras are predominately activated in high dynamic range (HDR) mode, however, legacy procedures to quantify image sensor noise were optimized for operation in standard dynamic range mode. This work discusses the theoretical background and the workflow of the photon-transfer curve (PTC) test. Afterwards, it presents example implementations of the PTC test and its derivatives according to legacy procedures and according to procedures that were optimized for image sensors in HDR mode.

Conventional image quality metrics (IQMs), such as PSNR and SSIM, are designed for perceptually uniform gamma-encoded pixel values and cannot be directly applied to perceptually non-uniform linear high-dynamic-range (HDR) colors. Similarly, most of the available datasets consist of standard-dynamic-range (SDR) images collected in standard and possibly uncontrolled viewing conditions. Popular pre-trained neural networks are likewise intended for SDR inputs, restricting their direct application to HDR content. On the other hand, training HDR models from scratch is challenging due to limited available HDR data. In this work, we explore more effective approaches for training deep learning-based models for image quality assessment (IQA) on HDR data. We leverage networks pre-trained on SDR data (source domain) and re-target these models to HDR (target domain) with additional fine-tuning and domain adaptation. We validate our methods on the available HDR IQA datasets, demonstrating that models trained with with our combined recipe outperform previous baselines, converge much quicker, and reliably generalize to HDR inputs.

This paper investigates the relationship between image quality and computer vision performance. Two image quality metrics, as defined in the IEEE P2020 draft Standard for Image quality in automotive systems, are used to determine the impact of image quality on object detection. The IQ metrics used are (i) Modulation Transfer function (MTF), the most commonly utilized metric for measuring the sharpness of a camera; and (ii) Modulation and Contrast Transfer Accuracy (CTA), a newly defined, state-of-the-art metric for measuring image contrast. The results show that the MTF and CTA of an optical system are impacted by ISP tuning. Some correlation is shown to exist between MTF and object detection (OD) performance. A trend of improved AP5095 as MTF50 increases is observed in some models. Scenes with similar CTA scores can have widely varying object detection performance. For this reason, CTA is shown to be limited in its ability to predict object detection performance. Gaussian noise and edge enhancement produce similar CTA scores but different AP5095 scores. The results suggest MTF is a better predictor of ML performance than CTA.

This paper presents the design of an accurate rain model for the commercially-available Anyverse automotive simulation environment. The model incorporates the physical properties of rain and a process to validate the model against real rain is proposed. Due to the high computational complexity of path tracing through a particle-based model, a second more computationally efficient model is also proposed. For the second model, the rain is modeled using a combination of a particle-based model and an attenuation field. The attenuation field is fine-tuned against the particle-only model to minimize the difference between the models.