In the same section
- Extension of GIS tool based on 360 multimodal image navigation
We develop internally a GIS project (web based), based on 3D scanning acquisitions, coupled with photographs. The aim of this project is the extension of the tool (e.g. for automatic imports, dynamic colorization).
- Deep neural network image processing
Please register to UV for PROJH419 for more details
contact:
Olivier.debeir@ulb.be
- Chest XR covid detection [ASSIGNED]
- Multi-camera pose estimation using MediaPipe [ASSIGNED]
Objet : Intégration d'un système de motion capture dans le Sam. Le Sam est un dispositif médical multisensoriel qui génère des environnements calculés en temps réel, basés sur les besoins émotionnels des patients. Afin de générer des feedbacks en fonction des mouvements du patient, un dispositif de computer vision va être intégré à l’architecture existante du Sam. Le système de capture doit répondre aux objectifs suivants : - Captation sur un seul individu à la fois. - Système capable de fonctionner avec une luminosité variable - Le patient est susceptible de tourner à 360°. Dans la mesure du possible, le système d'aquisition doit prendre en compte les angles morts. (Triangulation avec plusieurs caméras, par exemple). - Le degré de précision de la captation ne doit pas être 100% accurate. Le projet tourne sur le moteur de rendu 3D Unreal Engine 5 (c++). Le framework mediapipe semble le bon candidat pour l'aquisition des données, mais des questions subsistent concernant la capture de la profondeur et des angles morts.www.inmersiv.com - maxime@inmersiv.com
- Image acquisition with a plenoptic camera
PROJ-H402
Description: The project consists in the development of an acquisition software to save and process the images obtained by a plenoptic camera. Such a camera captures multiple micro-images in a single shot, allowing for 3d reconstruction (microscopy, industrial inspection, biomedical imaging, ...). Using the API provided for the camera, the student will develop the tools to view, save and process the images captured on the sensor. They will develop and optimize image processing algorithms, such as micro-image detection, over-exposure detection and image restoration.
Contact the project supervisor for more details.
Programming language: C,C++ or C#.
Supervisor: Sarah Fachada, sarah.fernandes.pinto.fachada@ulb.be
Director: Mehrdad Teratani, mehrdad.teratani@ulb.be
- Deep stereo techniques for dense multi-view images extracted from plenoptic camera
PROJ-H402
Description: This project aims to generate the depth maps of dense datasets such as a 5x5 multi-view image from a plenoptic camera using state-of-the-art deep learning stereo matching methods and understand their limitations. Plenoptic cameras can capture 3D light-fields thanks to their structure which has a micro-lens array between the main lens and the sensor plane. Middle-bury stereo datasets should be tested to check if the techniques are applicable. Finally, extracted plenoptic multi-view images will be tested.
Contact the project supervisor for more details.
Programming language: Python.
Supervisor: Hamed Razavi Khosroshahi, hamed.razavi.khosroshahi@ulb.be
Director: Mehrdad Teratani, mehrdad.teratani@ulb.be
- Graphical Interface for Fast Assessment of Multiple View Quality (GIFA-MVQ)
PROJ-H402
Description: While doing the research, researchers in our lab often must compare multiple sets of images. Even if some specific tools already exist, they don't cover all the use cases and can be cumbersome to use. To facilitate the comparisons, we would like for this project to have a nice graphical interface to conduct visual assessments between the datasets. Moreover, we would like this GUI to embed some useful tools in it (example: objective metrics).
For additional details, please contact the supervisor of the project.
Programming language: Preference for Python or C++ but another language can be used if there is a motivation for it (except javascript).
Supervisor: Laurie Van Bogaert, laurie.van.bogaert@ulb.be
Director: Mehrdad Teratani, mehrdad.teratani@ulb.be
- Pipeline for evaluation of lenslet video compression
PROJ-H402
Description : Lenslet images are images captured by a plenoptic camera that can be used in 3D and VR applications such as generating multiview images of a 3D scene. Lenslet images can be preprocessed in order to be more efficiently compressed. To test the efficiency of a preprocessing scheme and its impact on VR applications, several steps need to be performed such as preprocessing, compression, generating multiview images, and computing quality metrics. The aim of this project is to code a pipeline that includes all the steps necessary to test the performance of a preprocessing scheme.
Contact the project supervisor for more details.
Programming language: Python, C/C++
Supervisor: Eline Soetens, eline.soetens@ulb.be
Director: Mehrdad Teratani, mehrdad.teratani@ulb.be
- Real-time Multi-layer Rendering for OpenGL Applications
PROJ-H402
Description: Graphics libraries such as OpenGL and Unity popularized 2D immersive applications such as video games. Recently, developers published a framework called OpenXR to help them develop 3D immersive applications. While this framework is a great opportunity for Virtual Reality developers, it is mostly limited to Head-Mounted Displays. Therefore, in this project, we would like to develop a software that can connect an OpenGL application to our specific 3D display.
Contact the project supervisor for more details.
Programming language: C++, knowledge in graphics libraries as OpenGL is preferred
Supervisor: Armand Losfeld, armand.losfeld@ulb.be
Director: Mehrdad Teratani, mehrdad.teratani@ulb.be
- Automatic Light Field Content Generator
PROJ-H402
Description: In science, it is important to have several datasets with different configurations to understand the limitations and benefits of the researchers' method. In our lab, we deal often with light fields (i.e., a data format/specification that represents a scene) which are composed of numerous viewpoints of a scene that make them difficult to capture. Therefore, 3D softwares as Blender are used to produce synthetic datasets but generating manually different light fields is time-consuming. In this project, we propose to develop a script that generates randomly a light field based on a dataset of objects and textures. The software must be able to load meshes and textures from a collection, place the objects randomly or with a certain "human coherence", change their light behaviors, and render the light field.
Contact the project supervisor for more details.
Programming language: C++, knowledge in graphics libraries as OpenGL is preferred
Supervisor: Armand Losfeld, armand.losfeld@ulb.be
Director: Mehrdad Teratani, mehrdad.teratani@ulb.be