Interactive GUI Research Articles

The use of sound phonons in the development of NVH treatments for light commercial electric vehicles

Light Commercial Electric vehicles (LCEV) are becoming increasingly popular due to their environmental benefits and lower operating costs, making them a sustainable choice for businesses looking to reduce their carbon footprint and save fuel expenses. However, unlike Saloon and SUV electric vehicles LCEV's have to be potentially configured for a wide variety of operating roles including parcel delivery, mini-buses, retail wagons, ambulances and mobile workshops to name just a few. Consequently, when considering their refinement aspects during their development it has become economically restrictive to base this on actual physical prototypes. Acoustic modelling provides a suitable alternative, but any software package must be highly flexible so that multiple configurations can be examined without requiring time consuming re-modelling. This paper describes the use of sound phonons in an interactive GUI for the successful development process of an LCEV from target setting through to final optimised NVH package creation. The project involved full vehicle model creation via donor CAD, of a current ICE LCV, which was morphed into an EV to include an under-floor battery and rear drive system. Sub-system models of areas such as the interior bulkhead, battery pack insulation and EV encapsulation were also integrated during the optimisation process.

Spade: A Real-Time Fraud Detection Framework

In this demonstration, we introduce Spade, a sophisticated real-time fraud detection framework adept at navigating the complex transaction graph. Unlike conventional methods that are limited by performance and lack incremental update capabilities, Spade leverages advanced incremental updates in dense subgraph peeling algorithms to enhance efficiency, usability, and reduce latency, achieving a significantly better fraud prevention ratio. The demo showcases an interactive GUI prototype, allowing users to customize and explore dense subgraphs with various metrics and algorithms. This interactive demonstration also effectively highlights Spade's robust capacity to unearth fraudulent transactions within varied settings, including Grab's services and cryptocurrency transactions.

Adaptive MATLAB GUI for Hydraulic Brake Stopping Distance Simulation

The crucial segment of any vehicle is the braking unit which decides the vehicle’s performance and passenger safety on a large quotient. It is highly essential to decelerate the vehicle within a reduced distance to avoid any collision. The stopping distance is also dependent on many characteristics apart from the vehicle itself which are GVW(Gross Vehicle Weight), brake sizing, Road Friction, and other vehicle design parameters. Hereby, the result of any brake system is the Stopping Distance. Therefore, for every vehicle, it is inevitable to design in such a way that meets all requirements to pass the safety standards. In this article, we provide an interactive GUI of Mathematical modelling using MATLAB that depicts the running vehicle simulation while the brakes are applied and highlights the Stopping distance for any given inputs. Here, we have used a novel approach to calculate the stopping distance, given we have all vehicle properties. Importantly, cuts the design phase to a large extent saving an ample amount of time with all ease of access and dexterity. The core features of GUI: One will be able to generate and compare stopping distance and other brake-related results for any two given different design scenarios. GUI is automated so that it can simulate the stopping distance just by providing the brake/vehicle parameters inputs to the interface. Additional functionalities to compare design results for various standards (FMVSS, IS, ECE) can also be done effectively. Keywords: Braking Systems, MATLAB Application, Vehicle Safety, Stopping Distance

Campus Selection System

Abstract: The main aim of the Research Project is to develop software to manage university registration processes and create an interactive GUI through which TPO can manage all student details, messages on its console and send letters to students informing them about their student needs. registration process. Although the scope of this project is broad, the most important is the presentation of personal and educational information of students and companies. Students and companies can edit fields such as username and password. The project also allows the database administrator to access student and company information and put it into the main database. It is also possible to remove student and company information from the main database after the application is completed.

SACCHARIS v2: Streamlining Prediction of Carbohydrate-Active Enzyme Specificities Within Large Datasets.

Carbohydrates are chemically and structurally diverse, composed of a wide array of monosaccharides, stereochemical linkages, substituent groups, and intermolecular associations with other biological molecules. A large repertoire of carbohydrate-active enzymes (CAZymes) and enzymatic activities are required to form, dismantle, and metabolize these complex molecules. The software SACCHARIS (Sequence Analysis and Clustering of CarboHydrate Active enzymes for Rapid Informed prediction of Specificity) provides a rapid, easy-to-use pipeline for the prediction of potential CAZyme function in new datasets. We have updated SACCHARIS to (i) simplify its installation by re-writing in Python and packaging for Conda; (ii) enhance its usability through a new (optional) interactive GUI; and (iii) enable semi-automated annotation of phylogenetic tree output via a new R package or the commonly-used webserver iTOL. Significantly, SACCHARIS v2 has been developed with high-throughput omics in mind, with pipeline automation geared toward complex (meta)genome and (meta)transcriptome datasets to reveal the total CAZyme content ("CAZome") of an organism or community. Here, we outline the development and use of SACCHARIS v2 to discover and annotate CAZymes and provide insight into complex carbohydrate metabolisms in individual organisms and communities.

Refinement of Voting System through Visual Cryptography and Multi-factor Authentication to Further Mitigate Clone Phishing Attack

Background: The growing concern over phishing attacks on voting systems, particularly with the rise of online voting, has highlighted vulnerabilities in elections. The convenience of internet voting has led to security risks like clone phishing attacks. While online voting offers accessibility benefits, security, privacy, and usability issues have arisen. Multi-factor authentication (MFA) has been proposed to enhance security in mobile internet voting systems. Visual cryptography, dividing images into shares for decentralized data storage, is suggested to counter clone phishing. MFA's effectiveness in various sectors is established, and secure voting systems combining visual cryptography and blockchain have been proposed. This research sought to bolster the security of the voting system using visual cryptography and multi-factor authentication (MFA), with the goal of increasing voter trust and the reliability of the voting procedure, by designing a secure system and evaluating its performance, accuracy, and accessibility.
 Methodology: The researchers developed an improved voting system using visual cryptography and multi-factor authentication, assessed its performance against Eligo and Voxvote, utilized Java for programming, MYSQL via XAMPP for the database, HTML, CSS, JavaScript, and PHP (Laravel) for client-server sides. The Scrum agile methodology was followed, employing brief sprints for adaptive development. Evaluation was done through web tools and benchmarks. The system's architecture and flowchart were presented, featuring an interactive GUI, Java 2 platform, MySQL, PHP 7, and specific hardware/software requirements. System setup included database and server configuration, fingerprint scanner installation, and Java runtime setup. Deployment involved executing a built Java program, and system testing was conducted on Windows 10, utilizing the Apache server and initiating the "Electronic Voting" program for online multi-factor authentication.
 Results: Achieving a performance rate of 92%, the developed system surpassed its competitors Eligo and Voxvote, which achieved scores of 15% and 33% correspondingly, as indicated by the data. Eligo and Voxvote attained scores of 88% and 80% individually, whereas the newly designed system obtained a rating of 93% in terms of accessibility. Nonetheless, the research also highlighted specific downsides, encompassing intricacy, reluctance to embrace change, and technological barriers. To tackle these issues and enhance the adoption of such systems, these limitations underscore the necessity for further investigation and advancement.
 Conclusion: The study demonstrates that integrating multi-factor authentication and visual cryptography significantly enhances voting system security, reducing clone phishing risks. Visual cryptography secures decryption keys, preserving voting integrity, while multi-factor authentication adds defence against unauthorized access. The researcher's online voting system outperforms competitors in performance metrics and accessibility. The study underscores the importance of combining these techniques for improved voting system reliability and security.

KGNav: A Knowledge Graph Navigational Visual Query System

Visual query is a vital technique for comprehending and analyzing knowledge graphs, which provides an effective method to lower the barrier of querying knowledge graphs for non-professional users. Nevertheless, visual query techniques for knowledge graphs and ontologies that have emerged in recent years cannot bridge the gap between global information provided by the knowledge graph schema and underlying data of knowledge graph. Thus it cannot fully exploit the global information to navigate users for querying knowledge graphs. This demonstration showcases KGNav, a Knowledge Graph Navigational visual query system. KGNav (1) redefines the minimal unit of operation to abstract the conceptual hierarchy, i.e., Knowledge Graph Schema, in the domain from the original knowledge graph in an offline semi-automatic way through the equivalence relations between these units; it also (2) provides a series of operators and an interactive GUI to capture user query intentions, guiding users to explore the Knowledge Graph Schema to achieve in-depth analysis of knowledge graphs. We will demonstrate the capability of KGNav in reducing tedious queries, enabling users to swiftly grasp the structure of the knowledge graph, and performing queries through several fundamental scenarios.

A Review on Human–Computer Interaction Education Based on Active Learning using the DEMATEL Method

The objective of the study was to explore how web lectures could be used to enhance the learning experience in an introductory human-computer interaction course. The researchers introduced web lectures as a method to deliver lecture material prior to class, allowing more interactive learning activities during in-class sessions instead of passive listening. A quasi-experiment was conducted over a 15-week semester, involving 46 students who were divided into two sections of the same course. One section utilized web lectures, while the other followed traditional lectures. To ensure fairness, several control measures were implemented, including the same instructor for both sections and blind grading. The findings revealed that the section using web lectures achieved significantly better grades compared to the section with traditional lectures. Moreover, students in the web lecture section reported increasingly positive attitudes towards this approach throughout the course. This research provides a twofold contribution: firstly, it demonstrates a new way of utilizing existing technology to enhance learning outcomes, and secondly, it presents a comprehensive, quasi-experimental evaluation of web lectures in the specific context of the course. Various undergraduate courses incorporate the field of human-computer interaction (HCI) and emphasize a theoretical approach that encompasses the entire process of creating effective interfaces for end users. However, instructors encounter difficulties in keeping students engaged during the introduction of these concepts and encouraging practical experience in HCI. The aim of this study is to propose a teaching and learning method that enhances student retention and promotes a deeper understanding of HCI. This method consists of two components. The first part focuses on explaining the theoretical concepts of the discipline, while the second part involves implementing problem-based learning (PBL) within The DEMATEL method involves a systematic process of analyzing the relationships between different factors or variables in a system. It aims to uncover the cause-and-effect chains and understand the influence and impact each factor has on others. By employing DEMATEL, decision-makers can gain insights into the interdependencies among factors and determine which components play a significant role in shaping the system's behaviour. This method facilitates a comprehensive evaluation of the relationships, allowing for a more informed decision-making process. This model provides a clear representation of the complex system, highlighting the critical components and their influence on other factors. By visualizing and understanding the cause-effect chains, decision-makers can make informed choices and develop effective strategies for system improvement. Stability (ST), Speed (SP), System integration (SI), Aesthetic visualization (Av), Friendliness other learner (UF), interactive GUI(IG), interface flexibility (IF), User interaction with (IWS) Human–Computer Interaction Education Based on Active Learning Methods IWS is got the first rank whereas is the system integration is having the lowest rank. Human–Computer Interaction Education Based on Active Learning Methods IWS is got the first rank whereas is the system integration is having the lowest rank.

Nus-tool: A unified program for generating and analyzing sample schedules for nonuniformly sampled NMR experiments

Increases in digital resolution achieved by high-field NMR require increases in spectral width. Additionally, the ability to resolve two overlapping peaks requires a sufficiently long acquisition time. These constraints combine, so that achieving high resolution spectra on high-field magnets requires long experiment times when employing uniform sampling and Fourier Transform processing. These limitations may be addressed by using nonuniform sampling (NUS), but the complexity of the parameter space across the variety of available NUS schemes greatly hinders the establishment of optimal approaches and best practices. We address these challenges with nus-tool, which is a software package for generating and analyzing NUS schedules. The nus-tool software internally implements random sampling and exponentially biased sampling. Through pre-configured plug-ins, it also provides access to quantile sampling and Poisson gap sampling. The software computes the relative sensitivity, mean evolution time, point spread function, and peak-to-sidelobe ratio; all of which can be determined for a candidate sample schedule prior to running an experiment to verify expected sensitivity, resolution, and artifact suppression. The nus-tool package is freely available on the NMRbox platform through an interactive GUI and via the command line, which is especially useful for scripted workflows that investigate the effectiveness of various NUS schemes.

RoboDoc: Smart Robot Design Dealing with Contagious Patients for Essential Vitals Amid COVID-19 Pandemic

The COVID-19 pandemic took valuable lives all around the world. The virus was so contagious and lethal that some of the doctors who worked with COVID-19 patients either were seriously infected or died, even after using personal protective equipment. Therefore, the challenge was not only to help communities recover from the pandemic, but also to protect the healthcare staff/professionals. In this regard, this paper presents a comprehensive design of a customized pseudo-humanoid robot to specifically deal with contagious patients by taking basic vitals through a healthcare staff member from a remote location amid the COVID-19 pandemic. The proposed design consists of two portions: (1) a complete design of mechanical, electrical/electronic, mechatronic, control, and communication parts along with complete assembly to make a complete multitask-performing robot that interacts with patients to take vitals, termed as RoboDoc, and (2) the design of the healthcare staff side (master/operator side) control of a joystick mechanism with haptic feedback. The proposed RoboDoc design can be majorly divided into three parts: (1) the locomotion part is composed of two-wheeled DC motors on a rover base and two omni wheels to support the movements of the robot; (2) the interaction part consists of a single degree-of-freedom (s-DOF) neck to have communication with different heights of patients and (3) two anthropomorphic arms with three degrees-of-freedom (3-DOF). These parts help RoboDoc to reach to patient’s location and take all of the vitals using relevant devices such as an IR temperature thermometer, pulse oximeter, and electronic stethoscope for taking live auscultations from the lungs and heart of the patient. The mechanical design was created using solid works, and the electronic control design was made via proteus 8.9. For haptic teleoperation, an XBOX 360 controller based on wireless communication is used at the master/operator side. For the convenience of the healthcare staff (operator), an interactive desktop-based GUI was developed for live monitoring of all the vital signs of patients. For the remote conversation between the healthcare staff and the patient, a tablet is mounted (that also serves as the robot’s face), and that tablet is controlled via a mobile application. For visual aid, a DSLR camera is integrated and controlled remotely, which helps the doctor monitor the patient’s location as well as examine the patient’s throat. Finally, successful experimental results of basic vitals of the remote patient such as temperature sensing, pulse oximeter, and heart rate (using haptic feedback) were obtained to show the significance of the proposed cost-effective RoboDoc design.

WebArrayDB

Geospatial array DBMSs operate on georeferenced N -d arrays. They provide storage engines, query parsers, and processing capabilities as their core functionality. Traditionally, those have been too heavy for a Web browser to support. Hence, Web Applications, mostly Geographic Information Systems (GISs), run array management on their server back-ends that return small portions of the results via the network. We showcase WebArrayDB, the first geospatial array DBMS that can run completely inside a Web browser. We demonstrate that modern Web browsers, JavaScript, and respective software libraries are sufficiently mature to build and run such a feature-rich and powerful DBMS. A Web-based array DBMS should reduce server load, enable offline work, decrease network I/O, and improve user experience. We also present ArrayGIS, our new Web GIS based on WebArrayDB, and invite everyone to explore both via a freely accessible, informative, and interactive Web GUI.

Monitoring and Visualization of Crystallization Processes Using Electrical Resistance Tomography: CaCO3 and Sucrose Crystallization Case Studies.

In the current research work, electrical resistance tomography (ERT) was employed for monitoring and visualization of crystallization processes. A first-of-its-kind MATLAB-based interactive GUI application “ERT-Vis” is presented. Two case studies involving varied crystallization methods were undertaken. The experiments were designed and performed involving calcium carbonate reactive (precipitative) crystallization for the high conductivity solution-solute media, and the cooling crystallization of sucrose representing the lower conductivity solution–solute combination. The software successfully provided key insights regarding the process in both crystallization systems. It could detect and separate the solid concentration distributions in the low as well as high conductivity solutions using the visual analytics tools provided. The performance and utility of the software were studied using a software evaluation case study involving domain experts. Participant feedback indicated that ERT-Vis software helps by reconstructing images instantaneously, interactively visualizing, and evaluating the output of the crystallization process monitoring data.

Introduction of Lazy Luna an automatic software-driven multilevel comparison of ventricular function quantification in cardiovascular magnetic resonance imaging

Cardiovascular magnetic resonance imaging is the gold standard for cardiac function assessment. Quantification of clinical results (CR) requires precise segmentation. Clinicians statistically compare CRs to ensure reproducibility. Convolutional Neural Network developers compare their results via metrics. Aim: Introducing software capable of automatic multilevel comparison. A multilevel analysis covering segmentations and CRs builds on a generic software backend. Metrics and CRs are calculated with geometric accuracy. Segmentations and CRs are connected to track errors and their effects. An interactive GUI makes the software accessible to different users. The software’s multilevel comparison was tested on a use case based on cardiac function assessment. The software shows good reader agreement in CRs and segmentation metrics (Dice > 90%). Decomposing differences by cardiac position revealed excellent agreement in midventricular slices: > 90% but poorer segmentations in apical (> 71%) and basal slices (> 74%). Further decomposition by contour type locates the largest millilitre differences in the basal right cavity (> 3 ml). Visual inspection shows these differences being caused by different basal slice choices. The software illuminated reader differences on several levels. Producing spreadsheets and figures concerning metric values and CR differences was automated. A multilevel reader comparison is feasible and extendable to other cardiac structures in the future.

RestraintMaker: a graph-based approach to select distance restraints in free-energy calculations with dual topology

The calculation of relative binding free energies (RBFE) involves the choice of the end-state/system representation, of a sampling approach, and of a free-energy estimator. System representations are usually termed “single topology” or “dual topology”. As the terminology is often used ambiguously in the literature, a systematic categorization of the system representations is proposed here. In the dual-topology approach, the molecules are simulated as separate molecules. Such an approach is relatively easy to automate for high-throughput RBFE calculations compared to the single-topology approach. Distance restraints are commonly applied to prevent the molecules from drifting apart, thereby improving the sampling efficiency. In this study, we introduce the program RestraintMaker, which relies on a greedy algorithm to find (locally) optimal distance restraints between pairs of atoms based on geometric measures. The algorithm is further extended for multi-state methods such as enveloping distribution sampling (EDS) or multi-site lambda -dynamics. The performance of RestraintMaker is demonstrated for toy models and for the calculation of relative hydration free energies. The Python program can be used in script form or through an interactive GUI within PyMol. The selected distance restraints can be written out in GROMOS or GROMACS file formats. Additionally, the program provides a human-readable JSON format that can easily be parsed and processed further. The code of RestraintMaker is freely available on GitHub https://github.com/rinikerlab/restraintmaker.

Electrical impulse characterization along actin filaments in pathological conditions

We present an interactive Mathematica notebook that characterizes the electrical impulses along actin filaments in both muscle and non-muscle cells for a wide range of physiological and pathological conditions. The simplicity of the theoretical formulation, and high performance of the Mathematica software, enable the analysis of multiple conditions without computational restrictions. The program is based on a multi-scale (atomic → monomer → filament) approach capable of accounting for the atomistic details of a protein molecular structure, its biological environment, and their impact on the travel distance, velocity, and attenuation of monovalent ionic wave packets propagating along microfilaments. The interactive component allows investigators to choose the experimental conditions (intracellular Vs in vitro), nucleotide state (ATP Vs ADP), actin isoform (alpha, gamma, beta, and muscle or non-muscle cell), as well as a conformation model that covers a variety of mutants and wild-type (the control) actin filaments. We used the computational tool to analyze environmental changes such as temperature effects and pH changes of the surrounding solutions, as well as structural changes to an actin monomer due to radius changes. Additionally, we investigated for the first time the electrostatic consequences of actin mutations from different disease conditions. These studies may provide an unprecedented molecular understanding of why and how age, inheritance, and disease conditions induce dysfunctions in the biophysical mechanisms underlying the propagation of electrical signals along actin filaments. Program summaryProgram Title: EIAF (Electrical Impulse along Actin Filaments)CPC Library link to program files:https://doi.org/10.17632/k8vw4j6y75.1Developer's repository link:https://github.com/MarceloMarucho/SignalPropagationPathologicalConditionLicensing provisions: GPLv2Programming language: MathematicaNature of problem: Electrical properties of actin filaments in pathological conditions have yet to be studied, despite previous work showing influential consequences for physiological conditions on wild type actins.Solution method: The interactive Mathematica notebook is based on the Multi-scale Approach for electrical impulses along cytoskeletal filaments and uses non-linear considerations for the Boltzmann statistics of ionic distributions in the electrical double layer (EDL). Using the numerical methods incorporated in Mathematica’s default libraries, and an interactive GUI, the program allows users to conduct their own research by choosing the experimental conditions (intracellular Vs in vitro), nucleotide state (ATP Vs ADP), actin isoform (including muscle and non-muscles cells), as well as a conformation model that covers a variety of mutants and wild-type (the control) actin filament configurations.Additional comments including restrictions and unusual features: The program allows users to adjust pH values, input voltages, as well as study two default electrolyte conditions without rerunning the program. The program is fully editable, such that motivated users can change the default electrolytes to fit specific conditions in their research project in necessary.

Interactive GUI for the calculation of F obs–F obs electron-density difference maps and extrapolated structure factors based on the cctbx toolbox

Interactive GUI for the calculation of <i>F</i> _obs–<i>F</i> _obs electron-density difference maps and extrapolated structure factors based on the <i>cctbx</i> toolbox

Real-time monitoring and control of flow rate in transportation pipelines using Matlab-based interactive GUI and PID controller

&lt;p class="Normal1"&gt;In this paper, a Matlab based GUI and Propotinal Integral Dervative (PID) controller is designed to automatically regulate the flow-rate of the circulating fluid. When fluids are transported over long distances, the pressure and flow rate have to be monitored remotely in a control room. Using an HMI or Control Panels the flow rate can be increased or decreased to compensate for pressure drops or disturbances. This paper attempts to demonstrate such an Industrial Control Operation in a scaled-down environment. A Graphical User Interface or GUI is constructed which enables the Operator to monitor, as well as control an electronically actuated Control Valve which can efficiently regulate the flow-rate. Automatic operations have also been implemented using a PID controller algorithm, which tries to track the Set-point in Real-time.&lt;/p&gt;

Electromagnetic education: Development of an interactive GUI for demonstrating wave polarization

AbstractOver the years, we have seen rapid advances in technology that have contributed to the modern lifestyle. It is obvious that there is a strong correlation between technological advances and the quality of education that has been provided to students. It is fair to say that the current educational system is reasonably good, because it has successfully created a platform for innovation in science and technology. However, there seems to be a lack of integration of technology into the education system. The goal of this paper is to utilize technology to improve the quality of electromagnetics education. It documents the development of a MATLAB‐based graphical user interface to visualize plane wave polarization, electromagnetic wave propagation in two and three‐dimensions, plane wavefronts, and spherical wavefronts. The multiwindow app allows users to customize the wave parameters such as amplitude, frequency, phase shift, and attenuation constant. This interactive application does not aim to replace teachers and lecturers, but it serves to enhance the teaching process and to help students better understand the relevant concepts.

Optimizing AP and Beacon Placement in WiFi and BLE hybrid localization

Nowadays, wireless communication techniques, such as WiFi, Bluetooth low energy (BLE), etc., have been pervasive in our daily lives, and not only provide convenient data transmission services, but also enable popular indoor positioning and navigation services. The placement of wireless infrastructures like WiFi access points (APs) and BLE beacons have significant influences on the performance of localization. More importantly, since APs and beacons are probably used for both network access and localization, it is necessary to take into account both coverage and localization when deploying APs and beacons. In addition, there still exist other critical challenges, including optimizing AP and beacon placement in WiFi and BLE hybrid localization and optimizing the placement of extra APs and/or beacons in an existing wireless network, which are essentially NP-complete. This paper tackles these problems of optimizing AP and beacon placement by proposing a heuristic differential evolution algorithm based on the widely used Cramer-Rao lower bound (CRLB). To be specific, the CRLB is leveraged as a metric for localization and meanwhile a coverage degree criterion is defined as a metric for coverage, both of which are incorporated into the evaluation function of the differential evolution algorithm. Furthermore, instead of using the ideal log distance path loss (LDPL) model, the more practical Motley-Keenan model is adopted to reflect the influences of obstacles that are widespread in indoor environments. On these grounds, a software is designed and implemented based on Geotools to optimize AP and beacon placement in an interactive GUI manner. Finally, extensive simulations and field experiments are conducted, and a thorough comparison confirms the efficiency and effectiveness of the proposed algorithm.

Interactive Scene Generation via Scene Graphs with Attributes

We introduce a simple yet expressive image generation method. On the one hand, it does not require the user to paint the masks or define a bounding box of the various objects, since the model does it by itself. On the other hand, it supports defining a coarse location and size of each object. Based on this, we offer a simple, interactive GUI, that allows a layman user to generate diverse images effortlessly.From a technical perspective, we introduce a dual embedding of layout and appearance. In this scheme, the location, size, and appearance of an object can change independently of each other. This way, the model is able to generate innumerable images per scene graph, to better express the intention of the user.In comparison to previous work, we also offer better quality and higher resolution outputs. This is due to a superior architecture, which is based on a novel set of discriminators. Those discriminators better constrain the shape of the generated mask, as well as capturing the appearance encoding in a counterfactual way.Our code is publicly available at https://www.github.com/ashual/scene_generation.