Kurt Debattista

The University of Warwick · Warwick Manufacturing Group

Due to the costliness of labelled data in real-world applications, semi-supervised learning, underpinned by pseudo labelling, is an appealing solution. However, handling confusing samples is nontrivial: discarding valuable confusing samples would compromise the model generalisation while using them for training would exacerbate the issue of confirm...

Recently, Deep Learning-based methods for inverse tone mapping standard dynamic range (SDR) images to obtain high dynamic range (HDR) images have become very popular. These methods manage to fill over-exposed areas convincingly both in terms of details and dynamic range. To be effective, deep learning-based methods need to learn from large datasets...

Dense captioning creates diverse Region of Interests (RoIs) descriptions for complex visual scenes. While promising results have been obtained, several issues persist. In particular: 1) it is hard to find the optimal parameters for artificially designed modules (e.g., non-maximum suppression (NMS)) causing redundancies and fewer interactions to ben...

Camera image simulation is integral to the virtual validation of autonomous vehicles and robots that use visual perception to understand their environment. It also has applications in creating image datasets for training learning-based vision models. As camera image simulation takes into account a wide variety of external and internal parameters, a...

Precise scene understanding based on perception sensors' data is important for assisted and automated driving (AAD) functions, to enable accurate decision-making processes and safe navigation. Among various perception tasks using camera images (e.g. object detection, semantic segmentation), panoptic segmentation shows promising scene understanding...

A CF explainer identifies the minimum modifications in the input that would alter the model's output to its complement. In other words, a CF explainer computes the minimum modifications required to cross the model's decision boundary. Current deep generative CF models often work with user-selected features rather than focusing on the discriminative...

Virtual reality has great potential to enable remote collaborative work from anywhere in the world. Developing virtual reality into a platform suitable for natural interaction and immersive collaboration requires the experience to be reliably stable. For a networked collaborative environment, perceived smoothness of motion is limited by the tick ra...

Automated sorting through chemometric analysis of plastic spectral data could be a key strategy towards improving plastic waste management. Deep learning is a promising chemometric tool, but further development through multi-modal deep learning has been limited by lack of data availability. A new Multi-modal Plastic Spectral Database (MMPSD) consis...

The use of chemometric techniques with spectral data for sorting plastics to improve recycling rates have gained more attention in recent years. However, insufficient representation of polymer spectra in spectral databases has been one of the barriers to the further development of these techniques. This work aims to develop a polymer spectra datase...

Sheet metal stamping is widely used for high-volume production. Despite the wide adoption, it can lead to defects in the manufactured components, making their quality unacceptable. Because of the variety of defects that can occur on the final product, human inspectors are frequently employed to detect them. However, they can be unreliable and costl...

Dense captioning generates more detailed spoken descriptions for complex visual scenes. Despite several promising leads, existing methods still have two broad limitations: 1) The vast majority of prior arts only consider visual contextual clues during captioning but ignore potentially important textual context; 2) current imbalanced learning mechan...

Increasing plastic recycling rates is key to addressing plastic pollution. New technologies such as chemometric analysis of spectral data have shown great promises in improving the plastic sorting efficiency to boost recycling rates. In this work, a novel deep learning architecture, PolymerSpectraDecisionNet (PSDN) was developed, consisting of conv...

Detecting rare and costly defects, such as necks and splits in sheet metal stamping, remains challenging for deep learning models due to low failure rates entailing few available samples to train on. Synthetic images provide a simulated alternative; however, the two main current approaches have limitations for generating split defect images. Image...

Due to the costliness of labelled data in real-world applications, semi-supervised object detectors, underpinned by pseudo labelling, are appealing. However, handling confusing samples is nontrivial: discarding valuable confusing samples would compromise the model generalisation while using them for training would exacerbate the confirmation bias i...

High-end mobile devices now support displaying video in High Dynamic Range (HDR), delivering a significantly enhanced viewing experience over Standard Dynamic Range (SDR). However, more energy may be required to play HDR, impacting device battery life and reducing overall quality of experience. We present a new methodology for predicting the real-...

Deep learning-based semi-supervised learning (SSL) algorithms are promising in reducing the cost of manual annotation of clinicians by using unlabelled data, when developing medical image segmentation tools. However, to date, most existing semi-supervised learning (SSL) algorithms treat the labelled images and unlabelled images separately and ignor...

High dynamic range (HDR) displays based on liquid crystal panels require local dimming algorithms to reproduce content with high fidelity and HDR. However, most local dimming algorithms are developed by using hand-crafted features and most of them focus on low dynamic range images rather than HDR images. In addition, few local dimming algorithms ca...

Due to the costliness of labelled data in real-world applications, semi-supervised object detectors, underpinned by pseudo labelling, are appealing. However, handling confusing samples is nontrivial: discarding valuable confusing samples would compromise the model generalisation while using them for training would exacerbate the confirmation bias i...

High dynamic range (HDR) displays with dual-panels are one type of displays that can provide HDR content. These are composed of a white backlight panel and a colour LCD panel. Local dimming algorithms are used to control the backlight panel in order to reproduce content with high dynamic range and contrast at a high fidelity. However, existing loca...

Defect inspection is a crucial step in sheet metal stampingSheet metal stamping manufacturing. However, current inspection methods largely consist of visual inspection by trained operatives but are unreliable and prone to error. Computer vision techniques have the potential advantage of utilising low cost hardware to enable accurate classification...

Current appearance models for the sky are able to represent clear sky illumination to a high degree of accuracy. However, these models all lack a common feature of real-skies: clouds. These are an essential component for many applications which rely on realistic skies, such as image editing and synthesis. While clouds can be added to existing sky m...

Mismanagement of plastic waste globally has resulted in a multitude of environmental issues, which could be tackled by boosting plastic recycling rates. Chemometrics has emerged as a useful tool for boosting plastic recycling rates by automating the plastic sorting and recycling process. This paper will comprehensively review the recent works apply...

Sheet metal stamping is widely used for high-volume production. Despite the wide adoption, it can lead to defects in the manufactured components, making their quality unacceptable. Because of the variety of defects that can occur on the final product, human inspectors are frequently employed to detect them. However, they can be unreliable and costl...

Dense captioning provides detailed captions of complex visual scenes. While a number of successes have been achieved in recent years, there are still two broad limitations: 1) most existing methods adopt an encoder-decoder framework, where the contextual information is sequentially encoded using long short-term memory (LSTM). However, the forget ga...

This article proposes a Markov Chain Monte Carlo ( MCMC ) rendering algorithm based on a family of guided transition kernels. The kernels exploit properties of ensembles of light transport paths, which are distributed according to the lighting in the scene, and utilize this information to make informed decisions for guiding local path sampling. Cri...

Recently, Deep Learning-based methods for inverse tone-mapping standard dynamic range (SDR) images to obtain high dynamic range (HDR) images have become very popular. These methods manage to fill over-exposed areas convincingly both in terms of details and dynamic range. Typically, these methods, to be effective, need to learn from large datasets a...

The presence of people plays a key role in how an environment, both now and in the past, is perceived. While there have been numerous computer reconstructions of cultural heritage sites, many, if not most, of these do not contain people. One reason for this is the so called “uncanny valley” affect. Despite huge efforts to create highly realistic lo...

Inverse Tone Mapping (ITM) methods attempt to reconstruct High Dynamic Range (HDR) information from Low Dynamic Range (LDR) image content. The dynamic range of well-exposed areas must be expanded and any missing information due to over/under-exposure must be recovered (hallucinated). The majority of methods focus on the former and are relatively su...

Inverse Tone Mapping (ITM) methods attempt to reconstruct High Dynamic Range (HDR) information from Low Dynamic Range (LDR) image content. The dynamic range of well-exposed areas must be expanded and any missing information due to over/under-exposure must be recovered (hallucinated). The majority of methods focus on the former and are relatively su...

High dynamic range (HDR) technology allows more of the lighting in a specific scene to be captured at a set point in time, and thus is capable of delivering an overall view of the scene that more closely correlates with our visual experience in the real world, compared to standard, or low dynamic range (LDR) technology. Although HDR capabilities of...

Dual-panel displays require local dimming algorithms in order to reproduce content with high fidelity and high dynamic range. In this work, a novel deep learning based local dimming method is proposed for rendering HDR images on dual-panel HDR displays. The method uses a Convolutional Neural Network to predict backlight values, using as input the H...

Accurately identifying cloud types in images has multiple uses from meteorological science to computer graphics, especially as clouds are a major factor influencing atmospheric radiative transport. Understanding which cloud types are present in an image is typically performed on a coarse scale, where cloud types are identified per image, but do not...

Real-time high-fidelity rendering requires the use of expensive high-end hardware, even when rendering moderately complex scenes. Interactive streaming services and cloud gaming have somewhat mitigated the problem at the cost of response lag. In this paper we present ReGGI (Regular Grid Global Illumination), a distributed rendering pipeline that el...

Convolutional Neural Networks (CNNs) are the current de-facto approach used for many imaging tasks due to their high learning capacity as well as their architectural qualities. The ubiquitous UNet architecture provides an efficient and multi-scale solution that combines local and global information. Despite the success of UNet architectures, the us...

High dynamic range (HDR) technology enables the capture, storage, transmission and display of real-world lighting at a high precision as opposed to traditional low dynamic range (LDR) imaging. One of HDR’s main features is its ability to reproduce very bright and very dark areas simultaneously. Dynamic range describes the span between these extrema...

In order to enhance viewing experiences, a number of backlight local dimming (BLD) algorithms have been developed to improve the image contrast ratio and provide power efficiency for modern displays. In order to evaluate which BLD algorithm performs best for HDR images rendering on dual-panel displays, this paper presents a comprehensive subjective...

Virtual Environments (VEs) provide the opportunity to simulate a wide range of applications, from training to entertainment, in a safe and controlled manner. For applications which require realistic representations of real world environments, the VEs need to provide multiple, physically accurate sensory stimuli. However, simulating all the senses t...

Perceiving and understanding cyber-attacks can be a difficult task, and more effective techniques are needed to aid cyber-attack perception. Attack modelling techniques (AMTs) - such as attack graphs, attack trees and fault trees, are a popular method of mathematically and visually representing the sequence of events that lead to a successful cyber...

Serious games are accepted as an effective approach to deliver augmented feedback in motor (re-) learning processes. The multi-modal nature of the conventional computer games (e.g. audiovisual representation) plus the ability to interact via haptic-enabled inputs provides a more immersive experience. Thus, particular disciplines such as medical edu...

Computer-assisted multimodal training is an effective way of learning complex motor skills in various applications. In particular disciplines (eg. healthcare) incompetency in performing dexterous hands-on examinations (clinical palpation) may result in misdiagnosis of symptoms, serious injuries or even death. Furthermore, a high quality clinical ex...

Human facial recognition in the context of surveillance, forensics and photo-ID verification is a task for which accuracy is critical. Quite often limitations in the overall quality of facial images reduces individuals' ability in taking decisions regarding a person's identity. To verify the suitability of advanced imaging techniques to improve ind...

Interactive high-fidelity rendering is still unachievable on many consumer devices. Cloud gaming services have shown promise in delivering interactive graphics beyond the individual capabilities of user devices. However, a number of shortcomings are manifest in these systems: high network bandwidths are required for higher resolutions and input lag...

Real-time high-fidelity rendering requires the use of expensive high-end hardware, even when rendering moderately complex scenes. Interactive streaming services and cloud gaming have somewhat mitigated the problem at the cost of response lag. In this paper we present ReGGI (Regular Grid Global Illumination), a cloud-based distributed rendering pipe...

Consumer High Dynamic Range (HDR) displays are appearing on the market. Capable of generating a peak luminance of up to 2000nits, the improved dynamic range they provide can only be perceived when viewed in a dark environment. In this paper, we present a display architecture that is capable of generating a peak luminance of 10,000 nits. We demonstr...

High dynamic range (HDR) displays are capable of displaying a wider dynamic range of values than conventional displays. As HDR content becomes more ubiquitous, the use of these displays is likely to accelerate. As HDR displays can present a wider range of values, traditional strategies for mapping HDR content to low dynamic range (LDR) displays can...

Virtual Environments (VEs) provide the opportunity to simulate a wide range of applications, from training to entertainment, in a safe and controlled manner. For applications which require realistic representations of real world environments, the VEs need to provide multiple, physically accurate sensory stimuli. However, simulating all the senses t...

Attack graphs and attack trees are a popular method of mathematically and visually representing the sequence of events that lead to a successful cyber-attack. Despite their popularity, there is no standardised attack graph or attack tree visual syntax configuration,and more than seventy self-nominated attack graph and twenty attack tree configurati...

Recently, there has been significant progress in the research and development of High Dynamic Range (HDR) video technology and state-of-the-art video pipelines are able to offer a higher bit depth support to capture, store, encode and display HDR video content. In this paper, we introduce a novel HDR video compression algorithm which uses a percept...

High dynamic range (HDR) imaging provides the capability of handling real world lighting as opposed to the traditional low dynamic range (LDR) which struggles to accurately represent images with higher dynamic range. However, most imaging content is still available only in LDR. This paper presents a method for generating HDR content from LDR conten...

High dynamic range (HDR) imaging enables the full range of light in a scene to be captured, transmitted and displayed. However, uncompressed 32-bit HDR is four times larger than traditional low dynamic range (LDR) imagery. If HDR is to fulfil its potential for use in live broadcasts and interactive remote gaming, fast, efficient compression is nece...

An evaluation of a no-reference objective quality metric to accurately measure and compare the increased dynamic range capability of modern image capture systems is presented. The use of commercial off-the-shelf equipment and software allows manufacturers and users to produce results that are both transparent and comparable. The no-reference metric...

High dynamic range (HDR) imaging permits the capture, storage, display, and handling of real world lighting, removing the limitations that traditional imaging has that may lead to over- and underexposed pixels in images. In order to achieve this, HDR imagery requires the storage and manipulation of floating point data, which consumes more space tha...

Perceiving and understanding cyber-attacks can be a difficult task. This problem is widely recognised and well documented, and more effective techniques are needed to aid cyber-attack perception. Attack modelling techniques (AMTs) –such as attack graphs and fault trees, are useful visual aids that can aid cyber-attack perception; however, there is...

High dynamic range (HDR) imagery permits the manipulation of real-world data distinct from the limitations of the traditional, low dynamic range (LDR), content. The process of retargeting HDR content to traditional LDR imagery via tone mapping operators (TMOs) is useful for visualizing HDR content on traditional displays, supporting backwards-compa...

Virtual environments grant the ability to experience real-world scenarios, such as driving, in a virtual, safe and reproducible context. However, in order to achieve their full potential the fidelity of the virtual environment must provide confidence that it replicates the perception of the real-world experience. The computational cost of simulatin...

Maximizing performance for rendered content requires making compromises on quality parameters depending on the computational resources available . Yet, it is currently unclear which parameters best maximize perceived quality. This work investigates perceived quality across computational budgets for the primary spatiotemporal parameters of resolutio...

In the medieval period, Coventry, in the English Midlands was a major centre for tanning as well as for its better-known cloth manufacturing. Heavily damaged during the Second World War, and unsympathetically rebuilt in the 1950s and 1960s, there is little left in modern Coventry to remind visitors of this important period in Coventry's history. Th...

Accurate simulation of all the senses in virtual environments is a computationally expensive task. Visual saliency models have been used to improve computational performance for rendered content, but this is insufficient for multi-modal environments. This paper considers cross-modal perception and, in particular, if and how olfaction affects visual...

HDR video on mobile devices is in its infancy and there are no solutions yet that can achieve full HDR video reproduction due to computational power limitations. In this paper we present a novel and versatile solution that allows the delivery of HDR video on mobile devices by taking into account contextual information and retro-compatibility for de...

This book explores the methods needed for creating and manipulating HDR content. HDR is a step change from traditional imaging; more closely matching what we see with our eyes. In the years since the first edition of this book appeared, HDR has become much more widespread, moving from a research concept to a standard imaging method. This new editio...

Fidelity is of key importance if virtual environments are to be used as authentic representations of real environments. However, simulating the multitude of senses that comprise the human sensory system is computationally challenging. With limited computational resources, it is essential to distribute these carefully in order to simulate the most i...

Serious games are accepted as an effective approach to deliver augmented feedback in motor (re-)learning processes. The multi-modal nature of the conventional computer games (eg. audiovisual representation) plus the ability to interact via haptic enabled inputs provides a more immersive experience. Thus, particular disciplines such as medical educa...

The following special issue is dedicated to game technologies for serious applications. It is based on the best papers received for the Seventh International Conference on Virtual Worlds and Games for Serious Applications (VS-Games 2015) which took place in Skovde, Sweden, from 16 to 18 September 2015. The conference had 48 submissions from which t...

This paper presents the results of a user study which quantifies the relative and absolute quality of example-based texture synthesis algorithms. In order to allow such evaluation, a list of texture properties is compiled, and a minimal representative set of textures is selected to cover these. Six texture synthesis methods are compared against eac...

High Dynamic Range (HDR) video has emerged from research labs around the world and entered the realm of consumer electronics. The dynamic range that a human can see in a scene with minimal eye adaption (approximately 1,000,000: 1) is vastly greater than traditional imaging technology which can only capture about 8 f-stops (256: 1). HDR technology,...

High dynamic range (HDR) imaging has become an important topic in both academic and industrial domains. Nevertheless, the concept of dynamic range (DR), which underpins HDR, and the way it is measured are still not clearly understood. The current approach to measure DR results in a poor correlation with perceptual scores (r ≈ 0.6). In this paper, w...

The ability to capture digital engineering imagery during rocket launches in which all of the detail in the scene is present is critical for postlaunch processing and analysis. Rocket launches often present an extreme range of lighting during both day and night missions. Imaging rocket launches has two key challenges: capturing detail of the vehicl...

Mobile devices are now widespread and multimedia consumption on these devices has increased significantly in recent years. More and more high dynamic range (HDR) content is being produced and its imminent adoption by the broadcast community means that there will soon be a demand to visualize HDR content on mobile devices. Mobile devices, however, h...

For High Dynamic Range (HDR) content, the dynamic range of an image is an important characteristic in algorithm design and validation, analysis of aesthetic attributes and content selection. Traditionally, it has been computed as the ratio between the maximum and minimum pixel luminance, a purely objective measure; however, the human visual system'...

Rendering textures in real-time environments is a key task in computer graphics. This paper presents a new parallel patch-based method which allows repeatable sampling without cache, and does not create visual repetitions. Interchangeable patches of arbitrary shape are prepared in a preprocessing step, such that patches may lie over the boundary of...

Olfactory displays provide a means to reproduce olfactory stimuli for use in virtual environments. Many of the designs produced by researchers, strive to provide stimuli quickly to users and focus on improving usability and portability, yet concentrate less on providing high levels of accuracy to improve the fidelity of odour delivery. This paper p...

High fidelity rendering systems rely on accurate material representations to produce a realistic visual appearance. However, these accurate models can be slow to evaluate. This work presents an approach for approximating these high accuracy reflectance models with faster, less complicated functions in regions of an image which possess low visual im...

Applying image processing applications under complex or harsh lighting conditions can be a difficult challenge. In particular, face recognition can be prone to such limitations due to the uncontrolled nature of the applications to which it is applied. One of the conventional ways used to resolve this concern is by capturing images under controlled...