Skip to content

Ideas for Diploma Projects

Updated on 10.04.2024 (for 2025)

Overview

Below is a list of possible topics. These topics can be modified and adjusted at will. Any other similar topic is possible, after a discussion with me.

  • Applications in signal processing: image / video / audio / etc.
  • Embedded devices for acquisition of signals
  • Any topic that the student is interested in, in the domain of signal processing. Talk to me about it first.

Requirements for students

Students should have:

  • abilities and interest in programming: Matlab / Python / any other language
  • general knowledge in Digital Signal Processing and related domain

Topics

General Machine Learning algorithms

Outlier detection in network data using Python

Use algorithms for anomaly detection from the Python PyOD library to detect attacks from network traffic (e.g. detect network attacks, which are different from normal traffic data)

Reference:

  • https://pyod.readthedocs.io/en/latest/
  • also see the tutorials / examples there

Dataset:

  • https://research.unsw.edu.au/projects/unsw-nb15-dataset

Failure detection in industrial systems

Use algorithms for anomaly detection from the Python PyOD library to detect failures in industrial devices, like ball bearings, based on accelerometer data.

Reference:

  • https://www.mdpi.com/2227-9709/8/4/85
  • https://pyod.readthedocs.io/en/latest/

Other algorithms

Any other ideas or applications of Machine Learning algorithms from the PyOD and scikit-learn Python libraries are welcome.

AI and neural networks

Parking Spotter

Use object detection methods for detecting cars and free parking spots in a large parking area. Implement it live on a Raspberry PI or equivalent.

Example:

  • https://www.parking-spotter.com/

Parking Spotter

Car Distance Estimator

Use object detection methods for detecting the car in front of you and, based on the size of the car, estimate the distance to it. Implement it live on a Raspberry PI or equivalent.

Car Distance Estimator

People Counter

Implement an application to detect and count people from images/videos and report long-term statistics

Image/video processing

Lane detection

Use image processing algorithms to detect the lanes from a traffic video and understand:

  • which lane we are currently on
  • is there a curve next, or is the road straight

Implement it live on a Raspberry PI or equivalent.

Example:

  • https://medium.com/computer-car/udacity-self-driving-car-nanodegree-project-1-finding-lane-lines-9cd6a846c58c

Lane detection

Image inpainting using sparse representations

Image inpainting means „filling” some missing pixels or parts of an image. Implement one of the various methods for this, namely the one based on sparse decompositions of images in the frequency domain.

References:

  • https://github.com/ily-R/Image-Inpainting-via-Sparse-Representation

Image classification with SVM (or other) algorithms

Implement and evaluate an algorithm for image classification, aiming to detect if a person is or is not wearing a mask.

References:

  • https://ieeexplore.ieee.org/document/9301079

Methods for image denoising

Implement and compare a number of methods for reducing the noise in general images or other types of data (based on filtering, different transforms, spline smoothing etc).

Examples:

  • http://www.mickaeltits.be/open-image-restoration/

Object tracking methods in videos

Implement and evaluate one method of video object tracking. End goal: track cars in a traffic video.

References:

  • https://manivannan-ai.medium.com/multiple-object-tracking-algorithms-a01973272e52-

General signal processing with embedded devices

System for acquiring and processing ECG signals

Develop a system for acquiring and analyzing a person's ECG signal, based on an ECG sensor integrated with a Raspberry PI device.

References:

  • https://www.optimusdigital.ro/en/others/4411-sensor-module-puls-sparkfun-ad8232.html
  • https://www.instructables.com/id/Portable-ECG-Machine/
  • https://www.researchgate.net/publication/317368775_A_Portable_Real_Time_ECG_Device_for_Arrhythmia_Detection_Using_Raspberry_Pi