Student Projects

If you are interested in taking a project in our group, please contact the responsible person under “Detailed Description” of the project that you would like to choose.

PROJECT PROPOSALS 2018-2019


Point Cloud Compression

  • Deep neural network Based point cloud compression software development
    [Detailed Description]

    Point cloud imaging has recently emerged as a viable solution for immersive 3D content representation in augmented, mixed and virtual reality applications. The vast amount of data, though, needed to faithfully reproduce real-world sceneries with this type of imaging makes inevitable the demand for efficient compression solution.

    Visual data compression typically comes at the expense of distortions and the presence of artifacts that affect the visual quality of the compressed models. Thus, it is of crucial importance to not only reduce the amount of data needed to represent a model, but also to maintain the highest possible visual quality. The majority of point cloud compression schemes are currently based either on efficient geometrical data structures, or on projection-based solutions. Recently, deep convolutional neural networks have been proposed for point cloud compression purposes, showing remarkable performance gains with respect to the alternative solutions. Following the current trends, the objective of this project is to design and implement a deep neural network module for point cloud compression. The network should be able to efficiently compress point cloud models at various bitrates, while maintaining the highest possible visual quality.

    The following tasks should be performed during the project:

    • Study the state-of-the-art in deep neural network for point cloud representations.
    • Design and implement a deep neural network module for point cloud compression.
    • Identify suitable point cloud models for training.
    • Train the network and obtain the network parameters.
    • Quality assessment of the performance of the network.
    • Document all the development process and source code.

    Requirements: Requirements: Good background on image processing and machine learning. Good skills in programming.

    Contact: Evangelos Alexiou

    Group: Prof. Touradj Ebrahimi

    Suitable for: Bachelor Semester Project, Master Semester Project or Master Thesis in Electrical Engineering, Communication Systems, or Computer Science or equivalent

    Number of students: One

  • Perceptual-based point cloud compression
    [Detailed Description]

    With the increased interest of the market in adopting 3D imaging technologies, point clouds have emerged to practical content representations to faithfully reproduce real-life sceneries. However, the richer capabilities that are provided by this type of imaging come at the cost of huge amount of data that need to be stored, processed and delivered; thus, efficient compression solutions are inevitable. In fact, JPEG and MPEG standardization committees are currently active on this field, and the outcome of their efforts have shown remarkable progress [1].

    In this project the objective is to further improve coding efficiency by exploiting human visual attention. A data-set was recently released with fixation density maps of point cloud models [2], showing that human visual attention is attracted by low-level features, such as edges and contrast, and high level features, such as faces, for this type of imaging as well. Thus, one could exploit this information by assigning less bits to the regions of no interest, which would decrease the quality only on irrelevant parts of the content. In this case, the perceptual visual quality is expected to be the same, while less bits are being used. The student will have the opportunity to propose his/her own algorithm to improve current compression schemes, using the available data set, after getting familiarized with the state-of-the-art. Finally, the proposed solution will be benchmarked.

    The following tasks should be performed during the project:

    • Study and analyse the state-of-the-art in point cloud compression.
    • Acquire background knowledge in human visual attention.
    • Choose and learn in details a current compression algorithm, that will be used as a basis for the proposed algorithm.
    • Propose and implement an algorithm that takes under consideration regions of high interest.
    • Conduct a small-scale experiment to benchmark the proposed solution.
    • Document all the development process and source code.

    Requirements: Requirements: Good communication skills, good analytic and programming skills, solid mathematical background, image processing and compression is a plus.

    Contact: Evangelos Alexiou

    Group: Prof. Touradj Ebrahimi

    Suitable for: Master Semester Project or Master Thesis in Electrical Engineering, Communication Systems, or Computer Science or equivalent

    Number of students: One

    [1] S. Schwarz et al.: Emerging MPEG Standards for Point Cloud Compression IEEE Journal on Emerging and Selected Topics in Circuits and Systems, vol. 9, no. 1, pp. 133-148, March 2019

    [2] E. Alexiou; T. Ebrahimi: Towards modelling of visual saliency in point clouds for immersive applications. 26th IEEE International Conference on Image Processing (ICIP), Taipei, Taiwan, September 22-25, 2019

Image and Video Compression

  • Deep Neural Network Based Image Compression Software Development
    [Detailed Description]

    Image and video compression are essential tools for the storage and transmission of data for streaming and broadcasting. However, compression introduces artifacts and distortions in the original content, which may lead to a decrease in the quality of viewers’ experience. It is therefore important for content providers and service providers to be able to compress the content efficiently and achieve high quality using low bitrates. Deep convolutional neural networks have shown remarkable performance in image classification and object recognition. Following these breakthroughs, deep convolutional networks are increasingly used for other vision tasks including image compression. Current deep neural network based image compression algorithms present competitive results compared to state-of-the-art image compression methods such as JPEG.

    The goal of this project is to implement a deep neural network based image compression software. The network should be able to compress test images efficiently, achieving low bitrates while maintaining high visual quality. More specifically, the student should perform the following tasks:

    • Study state of the art on deep neural network based image compression,
    • Implement a deep neural network module that is able to carry out image compression,
    • Collect suitable images for training, with subjective ratings and create multiple datasets,
    • Train the network and obtain the network parameters,
    • Assess the performance of the network by performing objective tests on rate and distortion,
    • Document all the development process and final version of the code.

    Requirements: Requirements: Good background on image processing and machine learning. Good skills in programming.

    Contact: Pinar Akyazi

    Group: Prof. Touradj Ebrahimi

    Suitable for: Master Semester Project or Master Thesis in Electrical Engineering, Communication Systems, Computer Science, or equivalent.

    Number of students: One

  • Deep Neural Network Based Image Enhancement Software Development
    [Detailed Description]

    Image and video compression are essential tools for the storage and transmission of data for streaming and broadcasting. However, compression introduces artifacts and distortions in the original content, which may lead to a decrease in the quality of viewers’ experience. It is crucial for content providers and service providers to be able to deliver the content efficiently and achieve high quality using low bitrates. Deep convolutional neural networks have shown remarkable performance in image classification and object recognition. Following these breakthroughs, deep convolutional networks are increasingly used for other vision tasks including image enhancement and restoration. Current deep neural network based algorithms for removing compression artifacts present competitive results compared to state-of-the-art image restoration methodologies, both in terms of objective metrics and subjective quality.

    The goal of this project is to implement a deep neural network based image enhancement software. The network should be able to restore test images efficiently, achieving state-of-the-art objective results and consistent subjective ratings. More specifically, the student should perform the following tasks:

    • Study state of the art on deep neural network based image restoration,
    • Implement a deep neural network module that is able to carry out image restoration,
    • Collect suitable images for training, with subjective ratings and create multiple datasets,
    • Train the network and obtain the network parameters,
    • Assess the performance of the network by performing objective an subjective tests,
    • Document all the development process and final version of the code.

    Requirements: Requirements: Good background on image processing and machine learning. Good skills in programming.

    Contact: Pinar Akyazi

    Group: Prof. Touradj Ebrahimi

    Suitable for: Master Semester Project or Master Thesis in Electrical Engineering, Communication Systems, Computer Science, or equivalent.

    Number of students: One

Multimedia Quality Assessment

  • Objective quality metrics for assessment of compressed video
    [Detailed Description]

    Transmission and processing of video introduce artifacts and distortions in the original signal, which may lead to a reduction in the quality of experience. It is therefore important for content and service providers to evaluate the quality of video, using objective metrics and subjective quality assessment methodologies. Objective quality metrics employ mathematical models and numerically report the quality of content with respect to a set of given input parameters. Current state of the art has numerous objective quality metrics, while extensions and improvements to these models are still ongoing.

    The goal of this project is to implement state of the art quality metrics for assessment of compressed video. More specifically, the student should perform the following tasks:

    • Study state of the art objective quality metrics for video,
    • Implement software that is able to carry out objective quality assessment for video,
    • Collect suitable video sequences for experimentation and create a dataset,
    • Run experiments on video dataset,
    • Assess the performance of objective quality metrics implemented in the software by performing subjective tests.
    • Document all the development process and final version of the code.

    Requirements: Good skills in programming. Basic knowledge of image and video processing.

    Contact: Pinar Akyazi

    Group: Prof. Touradj Ebrahimi

    Suitable for: Bachelor or Master Semester Project in Electrical Engineering, Communication Systems, Computer Science, or equivalent.

    Number of students: One

  • Online subjective quality evaluation platform to assess quality of compressed video
    [Detailed Description]

    Transmission and processing of video introduce artifacts and distortions in the original signal, which may lead to a reduction in the quality of experience. It is therefore important for content and service providers to evaluate the quality of video, using objective metrics and subjective quality assessment methodologies. Subjective quality assessment methodologies employ human subject and report the quality of content with respect to viewers’ opinion. Procedures for subjective video quality evaluation involve psychophysical experiments, in which a number of viewers are given a set of stimuli to be consumed in either pre-defined laboratory settings or typical environments with less controllable conditions. The subjects’ ratings are recorded and reported as an indication of the quality of the video.

    In this project the goal is to implement an online subjective quality evaluation platform to assess quality of compressed video. Specifically, the student should perform the following tasks:

    • Study the state of the art and standards on subjective quality evaluation,
    • Implement an online subjective quality evaluation platform allowing users to vote on the quality of various content,
    • Collect suitable video sequences for experimentation and create a dataset,
    • Recruit users and perform subjective tests using the platform,
    • Use the ratings of the platform to assess the performance of objective quality evaluation metrics of choice.
    • Document all the development process and final version of the code and platform.

    Requirements: Good skills in web programming. Basic knowledge of image and video processing.

    Contact: Pinar Akyazi

    Group: Prof. Touradj Ebrahimi

    Suitable for: Bachelor or Master Semester Project in Electrical Engineering, Communication Systems, Computer Science, or equivalent.

    Number of students: One

  • Objective quality metrics for point clouds
    [Detailed Description]

    Recent trends show that 3D imaging technologies will dominate the market in the near future. Among the alternatives, point clouds denote a viable solution that has recently emerged for immersive content representation, proven by the current activities of JPEG and MPEG standardization committees. Yet, one of the open problems in this emerging field is the assessment of quality of models under typical degradations. In this project, the task is to investigate and propose new objective quality metrics for point cloud models.

    In essence, a point cloud can be defined as a collection of 3D points in space representing the external surface of an object. Each sample is defined by its position, while associated attributes may also be used in conjunction, in order to provide further information (e.g., color, normal vectors). The set of points that represent the 3D model can be interpreted as a (generally) irregularly sub-sampled surface. To quantify the degradation of a distorted model with respect to a reference, point-based [1], or projection-based [2] approaches can be used. In this project, we aim to investigate whether currently available metrics can be improved by enabling more advanced surface approximations, surface description quantities, and/or other domains to represent the signal (e.g., graphs) for the former class, or image processing tools for the latter.

    The following tasks should be performed during the project:

    • Study state-of-the-art in objective and subjective quality assessment of point clouds.
    • Analyse alternatives to improve performance of current solutions.
    • Propose and implement and algorithm.
    • Benchmark the proposed algorithm.
    • Document all the development process and source code.

    Requirements: Good communication skills, good analytic skills, solid mathematical background or knowledge of image processing

    Contact: Evangelos Alexiou

    Group: Prof. Touradj Ebrahimi

    Suitable for: Bachelor Semester Project, Master Semester Project, or Master Thesis in Electrical Engineering, Communication Systems, or Computer Science or equivalent

    Number of students: One

    [1] E. Alexiou; T. Ebrahimi: Benchmarking of Objective Quality Metrics for Colorless Point Clouds. 2018 Picture Coding Symposium (PCS), San Francisco, California, USA, June 24-27, 2018.

    [2] E. Torlig; E. Alexiou; T. A. Fonseca; R. L. de Queiroz; T. Ebrahimi: A novel methodology for quality assessment of voxelized point clouds. 2018. SPIE Optical Engineering + Applications, San Diego, California, USA, August 19-23, 2018. p. 107520I. DOI : 10.1117/12.2322741.

  • Subjective quality assessment of point clouds
    [Detailed Description]

    Point clouds denote a common and practical solution to capture, store, and deliver 3D visual information. This type of imaging has lately attracted a significant amount of interest, and is expected to be one of the dominant content representations in augmented reality, mixed reality and virtual reality applications that are emerging.

    Although a significant amount of work has been recently conducted [1], subjective quality assessment of point cloud representations are still an open problem. In particular, it is not clear whether variations of conventional methodologies proposed by various ITU recommendations, which are typically limited to passive evaluation, or more complicated evaluation scenarios that exploit the richer nature of the contents, are more suitable [2]. Moreover, the impact of different influencing factors that are introduced by different evaluation scenarios (e.g., type of rendering, interaction, etc.) in the perceptual visual quality, need to be analysed and quantified. The objective of this project is to get familiarized with the state-of-the-art in point cloud imaging and current solutions for content consumption. Subjective evaluation experiments using different frameworks will be conducted, and the results will be analysed to assess the impact of influencing factors that they introduce. The student will acquire knowledge on the state-of-the-art of point clouds, hands-on experience with multiple rendering prototypes and various 3D computer graphics frameworks (e.g., three.js, VTK, Unity). Moreover, the student will learn how to design subjective quality assessment experiments and apply statistical analysis tools on the collected data.

    The following tasks should be performed during the project:

    • Study and analyse the state-of-the art in point clouds.
    • Study and analyse subjective evaluation methodologies for quality assessment of 2D and 3D imaging.
    • Build background knowledge on point cloud rendering and get familiarized with rendering tools and development frameworks for 3D computer graphics.
    • Design and conduct large-scale experiment(s).
    • Statistical analysis of the collected data.
    • Document all the development process and source code.

    Requirements: Good communication skills, good analytic skills

    Contact: Evangelos Alexiou

    Group: Prof. Touradj Ebrahimi

    Suitable for: Bachelor Semester Project, Master Semester Project or Master Thesis in Electrical Engineering, Communication Systems, Computer Science or equivalent

    Number of students: One

  • Quality assessment of 3D content representations in Head Mounted Displays (HMDs)
    [Detailed Description]

    The current trend of adopting 3D technologies in imaging suggests that in the near future there will be a very substantial increase of new applications in virtual, augmented and mixed reality. In such applications, 3D visual data representations are used in order to provide enhanced experiences to the users. Among various alternatives, polygonal mesh and point clouds denote the most popular solutions.

    However, subjective and objective quality assessments for such 3D content representations are still an open problem. In this project, we aim to develop a tool in order to conduct subjective quality assessment in HMDs (e.g., HTC Vive). The software should be able to support both types of contents (i.e., point clouds and meshes) and the number of stimuli that could be visualized simultaneously by an observer. The student will be able to acquire hands-on experience on recent development tools and frameworks employed in virtual reality applications (e.g. Unity). In addition, he/she will have the opportunity to design his/her own experiment and acquire knowledge on the analysis of subjective data collected in the experiment.

    The following tasks should be performed during the project:

    • Build background knowledge by reviewing the state-of-the-art in point cloud imaging and mesh representation.
    • Study and analyze methodologies for subjective evaluation of quality of experience in multimedia.
    • Develop an application that allows users to interact with the content. The users should be able to evaluate provide their scores through the application.
    • Develop a tool that would allow to design new tests (add different contents, configure content features, add questions, keep log of data, etc.).
    • Conduct small-scale quality assessment experiments using the developed platform.
    • Document all the development process and all created codes.

    Requirements: Requirements: Good communication skills, good analytic and programming skills

    Contact: Evangelos Alexiou

    Group: Prof. Touradj Ebrahimi

    Suitable for: Bachelor Semester Project, Master Semester Project or Master Thesis in Electrical Engineering, Communication Systems, Computer Science or equivalent

    Number of students: One

Image privacy and security

  • Mobile App for Context-Aware Privacy Protection
    [Detailed Description]

    Wide spread of smart mobile devices with high-resolution cameras and user-friendly social networks make photo and video sharing an easy and popular activity. However, it also raises a wide concern on privacy issues as such content potentially reveals a lot of sensitive information about users. This project focuses on the research for solutions to privacy protection in online photo sharing systems in a context-aware manner. The goal of the project is to develop such a system that can help users automatically control the accesses to their online photos by others, depending on the image content and contextual information, e.g., viewer’s identity, location, time, etc.

    The following tasks should be performed during the project:

    • Research and analyze the existing context-aware privacy protection tools or systems, and the machine learning based solutions to privacy protection.
    • Research and analyze the modelling for contextual information and the features that impact privacy in the scenario of online photo sharing.
    • Based on our existing secure photo sharing application, ProShare, devise and develop an Android or iOS application that can collect users’ photos, contextual information, sharing decisions, etc.
    • Create a context-aware privacy database that contains user shared photos, contextual information (sharer and viewer’s), and corresponding sharing decisions, from a considerable number of subjects.
    • Based on the created database, train and test a privacy predictor for each subject to predict their privacy decisions, using available machine learning methods.

    Requirements: Basic knowledge of image processing, computer vision and machine learning; experience in mobile application development in iOS or Android. Experience with Facebook mobile API (Android or iOS) is a plus.

    Contact: msgto(‘touradj’, ‘ebrahimi’, ‘epfl.ch’)

    Group: Prof. Touradj Ebrahimi

    Suitable for: Master Semester Project, or Master Thesis in Electrical Engineering, Communication Systems, or Computer Science

    Number of students: one

  • Mobile App for Privacy Protection on iOS Platform
    [Detailed Description]

    Recently, public interest in privacy protection has increased dramatically. However, there is a general believe that protection of privacy will restrict online benefits of users. Therefore, protection of privacy in such a way that does not distract online habits of people is needed. This project focuses on visual privacy protection in images. Specifically, the intention of the project is to develop a mobile (iOS platform) application that would be able to obfuscate personal visual information in an image in a secure and recoverable way and share images via online social networks in a secure way.

    The following tasks should be performed during the project:

    • Research and review the existing visual privacy protection tools, as well as the way to share and manage secure content in social networks.
    • Design an app on smartphone with iOS operating system. An iPhone will be provided by the lab.
    • Minimal requirements of the app include:
        • Implementation of security processing (e.g. scrambling) for images on the mobile side.
        • Multi-region processing on image using touch screen.
    • Implementation of a simple key management system.

    Requirements: Basic knowledge of image processing, good programming skills in Objective-C, experience in iOS deveopment.

    Contact: Touradj Ebrahimi

    Group: Prof. Touradj Ebrahimi

    Suitable for: Bachelor Project or Master Semester Project in Electrical Engineering, Communication Systems, or Computer Science

    Number of students: one

  • Mobile App for Privacy Protection on Android Platform
    [Detailed Description]

    Recently, public interest in privacy protection has increased dramatically. However, there is a general believe that protection of privacy will restrict online benefits of users. Therefore, protection of privacy in such a way that does not distract online habits of people is needed. This project focuses on visual privacy protection in images. Specifically, the intention of the project is to develop a mobile (Android platform) application that would be able to obfuscate personal visual information in an image in a secure and recoverable way and share images via online social networks in a secure way.

    The following tasks should be performed during the project:

    • Research and review the existing visual privacy protection tools, as well as the way to share and manage secure content in social networks.
    • Design an app on smartphone with Android operating system. An Android phone will be provided by the lab.
    • Minimal requirements of the app include:
        • Implementation of security processing (e.g. scrambling) for images on the mobile side.
        • Multi-region processing on image using touch screen.
    • Implementation of a simple key management system.

    Requirements: Basic knowledge of image processing, good programming skills in Java, experience in Android deveopment.

    Contact: Touradj Ebrahimi

    Group: Prof. Touradj Ebrahimi

    Suitable for: Bachelor Project or Master Semester Project in Electrical Engineering, Communication Systems, or Computer Science

    Number of students: one

  • Privacy Preserving Photo-Sharing Application
    [Detailed Description]

    The rapid growth of photo sharing through social media raises serious questions related to ownership, privacy and access to shared images. From the user perspective, effective privacy protection tends to impose restrictions on how users share and access pictures, making privacy protection unattractive. To address such issues, MMSPG has developed ProShare, a mobile App through which pictures can be protected, shared and made selectively accessible in a transparent manner while incurring minimal distraction to the user. To date a large effort has been invested in the development and implementation of the ProShare mobile App. Less attention has been directed at the server side realization of ProShare.

    The objective of this student project is to enhance the server side implementation of the ProShare service.

    The following tasks should be performed during the project:

    • Study and understand the ProShare service and its implementation (both server side and client side)
    • Review server side architecture and compare this to state-of-the-art implementations for similar services
    • Propose modifications and enhancements to the existing server side implementation
    • Implement a migration process allowing to move the ProShare server to a new computing platform
    • Implement robust and reliable session management
    • Propose and implement additional service features
    • Implement back-end tools for the analysis of usage and user statistics

    Requirements: Good communications skills. Good understanding of web server technologies including Apache, MySQL and PHP. Good abilities to think at the systems level.

    Contact: Touradj Ebrahimi

    Group: Prof. Touradj Ebrahimi

    Suitable for: Bachelor Semester Project in Computer Science, Communications Systems, Electrical Engineering or equivalent.

    Number of students: One

Eye-tracking and Visual attention

  • Modeling visual attention of 3D models in virtual reality
    [Detailed Description]

    Virtual reality (VR) applications aim to faithfully reproduce real-world sceneries allowing high levels of immersion and engagement of the user with the virtual content. For specialized VR applications, for example in usability or psychology research, among others, gaze information can be very useful. For instance, by analyzing recorded gaze data, researchers can draw conclusions concerning the way users consume and behave with the contents under inspection.

    In this project, the objective is to get familiarized and use an eye-tracking device in order to record and collect gaze information from humans consuming 3D models in a virtual environment. The virtual scene will be designed and developed using a real-time 3D development platform (e.g., Unity). The student will acquire knowledge on human visual attention basics, design and implementation of virtual scene(s) in a modern 3D development engine, hands-on experience with a state-of-the-art consumer market eye-tracking device, post-processing and analysis of the recorded user behavioral data. The reader could refer to [1] for the illustration of a similar study.

    The following tasks should be performed during the project:

    • Acquire knowledge on human visual attention.
    • Develop virtual scene(s) to display 3D models in a development platform of preference.
    • Get familiarized with the eye-tracking device and the API provided by the platform of preference.
    • Conduct a visual attention experiment to collect gaze information.
    • Post-process and analyse the recorded human behavioral data.
    • Document all the development process and source code.

    Requirements: Requirements: Good communication skills, good analytic and programing skills. Knowledge of computer graphics and/or familiarization with game development engines is a plus.

    Contact: Evangelos Alexiou

    Group: Prof. Touradj Ebrahimi

    Suitable for: Bachelor Semester Project, Master Semester Project, or Master Thesis in Electrical Engineering, Communication Systems, Computer Science or equivalent

    Number of students: One

    [1] E. Alexiou; T. Ebrahimi: Towards modelling of visual saliency in point clouds for immersive applications. 26th IEEE International Conference on Image Processing (ICIP), Taipei, Taiwan, September 22-25, 2019