ľ�ϸ���Ӱ��

See below an overview of the Honors Research Projects of the Bachelor's of Informatiekunde (INKU) and Informatica (INCA).

Continuum: a Three-Dimensional Continuous Space Microbial Simulation Framework

Continuum is a microbial simulation framework that helps biologists with prior coding experience quickly create fast simulations. It is a set of libraries written in C# that provide the basic tools for making a simulation, while leaving any simulation specific details to the user.

Simulations are run in 3D in continuous space to allow for a high freedom of simulation type. The provides acceleration structures allow for fast processing of collision checks, nearest neighbour searches, and range queries. The framework also provides support for writing to disk and to load simulations form disk.

Continuum comes packaged with an additional library for running simulations in different use cases. Such as a 3D and 2D renderer to help with debugging and visualization. And a console application to allow for maximum speed while running simulations.

All of this is provided along with the promise that simulations made with Continuum remain highly customizable as almost all of the core simulation logic can be adjusted to fit the needs of the user. Furthermore, the project comes with a detailed user guide as well as theoretical background information to ensure users can make optimal choices when designing their simulations.

Year: 2024/2025
Students: Philip Tap, Karel Boerstoel, and Kaan Pagie
Supervisors: Bram van Dijk

Optimal Pixel Cropping Layouts

Our main goal in this research was to look into how to find optimal layouts for a pixel cropping grid. During the first half of the project, we carried out a statistical analysis of the experiments by investigating whether neighboring crop types influenced the fresh marketable weight, and found no significant intercrop interactions. The second half of the project was spent developing a program that generates an optimal pixel farm layout, assuming known intercrop relations. We first attempted this by developing a MaxSAT based solution, but then found that this would not be fast enough for our purpose. We then attempted a MIQP based solution, which often finds very good solutions early in the solving process but then spends a lot of time proving its optimality. We suspect that more sophisticated heuristic algorithms can give good approximately optimal solutions.

Our research has contributed to the research of the ‘Going Off The Grid’ team from the Centre of Unusual Collaborations of the EWUU alliance. In this alliance, there are many researchers from different disciplines that can use our study on ideal pixel cropping layouts to continue their research.

Year: 2024/2025
Students: Wieke Boerop, Siem van Opstal, Emanuel Rietveld
Supervisors: Maarten Löffler

Musify

The goal for this project was to create a music recommender which could more closely fit the users needs. This is done by adding more user control in the form of various filters. The original recommendations for the users are done by comparing the user’s 50 most listened to songs in the last six months to a big dataset and finding the most compatible songs through a cosine similarity function. Once the recommendations are made users can save the entire list of recommendations to a spotify playlist directly or, if they only want certain songs in there, they can select the songs they like and get a playlist with only the selected songs.

Year: 2024/2025
Students: Samuel Eikelenboom, Jasper Alkema, Milan van der Burg
Supervisors: Karlijn Dinnisen, Hanna Hauptmann

Seeds of Knowledge

This project focused on building a custom digital library for Seeds Afghanistan, a research-based NGO that supports food security in rural parts of the country. Many important reports from organisations like the UN and WFP are hard to find or spread across different platforms. We created a web-based content management system that collects these reports in one place. Users can search for reports by keyword and filter by organization or date and administrators can upload new reports and edit report details. The system was designed to be easy to manage without technical knowledge. Our goal was to make it easier for local decision-makers and aid workers to access the information they need.

Year: 2024/2025
Students:  Emma den Hartogh and Mariëtte Smits
Supervisors:  Jaco Smit, Ioanna Lykourentzou and Mihaela Mitici

Robot framework

By default, the only applications for this robot could be made in the included “kiosk” application. This environment is insufficient for complex applications, as it only allows very limited programming, while the display was constrained to one specific layout. The original goal of creating a “receptionist robot” cannot be achieved with this. We have made it possible to control the robot by creating an Android app, since the robot runs on Android. In the process, we created a relatively easy-to-use API that handles the most common features. This comes with a demo application which implements controlling the robot from a game controller, simple movement, navigation to a point in the room, and navigation to the charger. The project comes with documentation for the API, as well as more general documentation about how we got it working and how you can reproduce our steps.

See also the following GitLab links:

• (not by us, included just for reference)

Year: 2024/2025
Students: Stefan Huneker, Robin Naafs, Daan van der Putten
Supervisors: Maartje de Graaf

Contact Information System (CIS)

This project addresses two issues: connecting your social media contacts to your address book, and keeping track of events you have attended with others. The Contact Information System (CIS) addresses both of these issues by linking contact information, and recording events in a separate program.

Year: 2023-2024
Students: Yorick Spekle (INCA)
Supervisors: Yannis Velegrakis

Dream Team (2022-2023 and 2023-2024)

What are the ingredients of a good student team? Oftentimes, university students need to work in teams, for example for project-based courses. The success of a course or study team depends on multiple factors, such as interpersonal compatibility, availability, compatible personalities or working styles. Teachers may also have their own requirements, for example regarding the skillset that the members of a course team should cover.

For the first part of this project (2022-2023) the students designed an online tool that helps teachers place students into teams in such a way so as to guarantee both the students’ preferences, and the lecturer requirements. In the project’s second part (2023-2024) the students focused on improving the process of forming teams, including improving functionality for the teachers, and improving the installation process and documentation.

The work is part of a larger project on computational methods for team formation by the

Year: 2022-2023 and 2023-2024
Team members: Wessel Boers (INKU, 2022-2023), Sophia Lin (INCA, 2022-2023, and 2023-2024), Leon Verhaegh (INKU, 2022-2023), Olivier van Mierlo (INKU, 2023-2024), Tygo Mourits (INCA, 2023-2024)
Supervisor: Ioanna Lykourentzou

Historic Utrecht: An Augmented Reality Soundscape Experience

This project seeks to develop a new form of storytelling that offers an immersive experience, similar to the “in the flow” feeling one gets when reading a book.

The application uses Augmented Reality (AR) in the form of soundscapes, which are carefully selected sounds that improve the outdoor walking experience. The project focuses on cultural heritage, particularly the city of Utrecht. Selected soundscapes, combined with AR, guide users as they walk past historical landmarks. This approach enables users to experience the city in a unique way, helping them forget about their modern surroundings, and immerse themselves in the city’s history.

Year: 2023-2024
Students: Sven Lakerveld (INKU)
Supervisors: Ioanna Lykourentzou

Open Science: Predictive Maintenance

The transition of research and its tools to Open Science is vital for collaboration within science, yet many see it as a necessary evil and struggle to incorporate open principles in their research.

In this project, we initially collaborated with an assistant professor on her research project designed to dynamically predict the Remaining Useful Life (RUL) of aircraft engines, adapting it to meet open science standards. To realise this, we extensively refactored and documented the code base, focusing on modular code structures and easy-to-understand documentation.

After this initial collaboration, we worked alongside Ioanna Lykourentzou to write a “12 Days to Open Software” case study to assist researchers in adapting their own code to open science standards. This way, we can collectively begin to build the bridges of communication between disciplines.

Year: 2023-2024
Students: Mathieu Chappin (INCA), Robin Schenkels (INCA)
Supervisors: Mihaela Mitici

Sketch to Success: A student-led INCA/INKU workshop at the Honors Community Day

The honours students of the Faculty of Computer & Information Sciences organized the Sketch to Success workshop as part of the Honours Community Day on 16th May 2024. In this workshop, they dived into project development and product improvement cycle, featuring a great talk by Dr. Gerard Wagenaar titled: 'Agile Software Development: WYSIWYG?'. They also showcased the projects that they have worked on over the year, and highlighted how they went from initial sketches to tangible projects. The project provided insights into the world of usability for a broad audience of developers, designers and end-users.

Year: 2023-2024
Students: Ruben de Boer (INCA), Robin Schenkels (INCA), Mathieu Chappin (INCA),  Yorick Spekle (INCA), Tygo  Mourits (INCA), Olivier van Mierlo (INKU), Guus van Hout (INKU), Sven Lakerveld (INKU)
Supervisors: Zerrin Yumak and Ioanna Lykourentzou

ToothSnap

In this project, we developed ToothSnap, an innovative application that educates users on how to improve their dental health. Its user-friendly and educational nature provides an engaging and interactive experience for learning and improving dental health.

The application uses brush patterns from disclosing tablets combined with a machine learning model to provide personalised brushing technique feedback. This interactive approach helps users visually understand areas they may have overlooked, or need to focus on, promoting better oral hygiene practices. The app provides educational tips for proper brushing techniques, as well as personalised advice based on the user’s brushing habits.

Year: 2023-2024
Students: Guus van Hout (INKU)
Supervisors: Ioanna Lykourentzou

Virtual Human Chatbot Project

In this project, we created a realistic 3D virtual human chatbot capable of answering questions about one’s own dataset. Specifically, the bot can respond to questions about the Game and Media Technology (GMT) master program. It is intended to help prospective GMT master students find information quickly in an engaging and fun way; for example, it could be used at an open day to pique people’s interest in the master’s program or on the university’s website to assist students in finding answers to their questions.

The bot speaks in a natural dialogue flow with mixed initiative, giving the feel of conversing with a real human. In order to appear as natural as possible, it is able to blink and make facial expressions. The assistant uses Unreal Engine’s MetaHumans for the 3D human, Nvidia’s Audio2Face for lip-sync, OpenAI’s GPT4o for chat, and Langchain to feed GPT4o with data.

Year: 2023-2024
Students: Ruben de Boer (INCA), Pascal Tollens (GMT Master student)
Supervisors: Zerrin Yumak

Big Data History of Music

In today's music landscape, the abundance of music sources can lead to confusion for enthusiasts. Conflicting information across various platforms can leave music fans perplexed. This project delves into the creation of a personalized music service designed to streamline the experience. Big Music consolidates data from multiple open music databases into a cohesive platform that allows users to extensively explore artists and their music. Through a web application that can connect with the APIs of various music providers, such as Spotify and MusicBrainz, users gain access to a unified interface that simplifies music discovery and management and permits obtaining extensive information about artists.

Year: 2022-2023
Team members: Damian Schilder (INKU), Eise de Vries (INKU), Fatih Enkür (INKU)
Supervisor: Frans Wiering

Dutch Natural Language Processing for clinical text anonymization

To use clinical texts for scientific research, anonymization of personal sensitive information is required by data protection regulations. Multiple methods exist to automatically perform this task. In this project, the performance of the rule-based Deduce algorithm was compared with that of the neural-based Deidentify algorithm.  As a preparation to the comparison, an annotated train-test dataset needs to be established. This corpus will form the ‘golden standard’. A first program was written that enabled annotating clinical texts for this purpose.

The Deduce algorithm is ready to go, but the Deidentify algorithm first needs training on a subset of the annotated texts. To accomplish this, a second program was written. For the actual comparison, a third program was written, where both algorithms perform their anonymisation task on the same test data. Afterwards, performance can be evaluated against the ‘golden standard’, for which a fourth program was written. Overall, this project covered all the necessary preparations to carry out the comparison, through which a full evaluation can take place using UMC Utrecht texts.

Year: 2022-2023
Team members: Collin C. Happé (INKU)
Supervisor: Pablo Mosteiro Romero

Raytracer

In this project, the students developed a modular didactic Raytracer to educate users on ray tracing fundamentals. Its modularity allows for layered learning, with each feature serving as a building block, enhancing comprehension. The developed Raytracer covers various techniques, from Whitted-style to path tracing and spectral ray tracing. It supports important features like Bounding Volume Hierarchy, 3D Meshes, and diverse materials (glass, metals, plastics). Additionally, it offers an accurate camera simulation with depth of field and other optical phenomena. The configurable nature of these features enables students to interactively manipulate and understand the complexities of ray tracing. This educational tool provides an engaging and interactive experience for learning ray tracing concepts.

Year: 2022-2023
Students: Thijs Boerefijn (INCA), Joachim Dekker (INCA), Gabriel Gonzalez (INCA), Guido Mintjes (INCA) and Joen van de Vorle (INCA)
Supervisor: Peter Vangorp

Data visualization challenge: Visualizing an Opera Network through Time

For this project, the students joined forces to address a data visualization challenge, namely effectively illustrating an opera network, using data from opera performances that took place across Europe between 1775 and 1833. The outcome of this project was submitted to the , which was held in conjunction with the .

Year: 2021-2022
Team members: Harpreet Kaur (INKU), Hedde van Heerde (INCA), Thomas Hietkamp (INKU)
Supervisor: Maarten Löffler, Ioanna Lykourentzou

Holland’s best liar: Game app for detecting lies

This project initiated the development of BestLiar, a citizen science game, the aim of which is to collect data on the lying behavior of individuals and how well others are at detecting these lies. The app includes different modes, including the ability of users to record a short clip of themselves lying, or stating whether the narrator is telling the truth or lying, or collaborative games that can be played with friends or family. On the technical side, the game was developed in Unity, and included API call connections to database and back-end services. The app is now available for download in the .

Year: 2021-2022
Team members: Bastiaan Van Gilst (INKU), Florence Herrent (INKU), Oliver Janssen (INKU), Björn Koemans (INKU), Xander Poortvliet (INKU)
Supervisor: Ronald Poppe

Explanations for the Dutch National Police: a Theoretical Investigation

One task of the police is the intake of citizens’ reports on crimes: the citizen tells the police what happened; subsequently, additional questions can be asked to determine if the citizen has been the victim or witness of a crime. Certain high-volume crimes can be reported online. This can be as simple as filling out a web form, but can also be a more involved online dialogue with a (possibly artificial) agent. If the reporting can be done online through an artificial agent, formal argumentation is used to derive conclusions, such as whether the citizen's situation is a case of fraud. Once the report is complete, the system will return its conclusion and the citizen might wish to request an explanation. This project explored the extraction of explanations from formal argumentation frameworks (which can be understood as directed graphs), focusing on understanding what types of explanations can be derived from such frameworks and what type is best suitable to return to a human user.

Year: 2020-2021
Team members: Jeroen Lombaers
Supervisor: AnneMarrie Borg

Open Source Data Diode

In collaboration with The Hague Security Delta, this project contributed to providing advice regarding the governance structure of Open-Source Data Diode: a combination of open hardware and open source software solutions that enables one-direction information traffic transmission. The advice, including recommendations on decision-making, financing, licensing, and legal issues, contributes to improved policies regarding citizen defense and open-source models.

Year: 2019-2020
Team members: Ikram Benhaddou (INKU), Luc Engelhard (INKU), Bart van As (INKU)
Supervisor: Slinger Jansen

Sounds Good: a playlist can change your life

Music plays a significant role in our daily lives. Some individuals enjoy listening to music while working out, while others find that they can enhance their performance and concentration by listening to their favorite songs. This project aimed to explore the impact of playlists on motivation and focus, with particular emphasis on the concept of flow. Described in the self-help book "Your Playlist Can Change Your Life," flow is a mental state characterized by complete attention to a task, leading to a loss of self-awareness and perception of time. The project focused on scientifically validating the propositions of the aforementioned book through user experimental evaluations, specifically concerning the creation of playlists that improve flow. Additionally, a digital prototype was developed targeting college students, aiming to assist them in enhancing task focus, minimizing distractions, and cultivating a sense of flow while studying through music.

Year: 2019-2020
Team members: Inge van Dijk (INKU), Thirza Hiwat (INKU), Tim van Kemenade (INCA), David Schouten (INKU)
Supervisor: Robbert-Jan Beun

Educational Computer Science Escape Room

This project focused on designing and building an educational computer science Escape Room. The purpose of the room is to engage students in computer science topics in a playful manner. At the same time, effort was placed to ensure that the teacher can easily set up the room in a low-cost manner, contributing to computer science education democratization. Along with a functioning prototype built with HTML, CSS, Javascript and PHP, the project outputs include the escape room’s puzzles, solutions and a ‘script’ to explain the flow of the room.

Year: 2018-2019
Team members: Sytze Simonse
Supervisor: Ioanna Lykourentzou