idigitall, my, digital, everything, blog, student, technology, technologist, futurist, software engineering, south africa, africa, gabor, support vector machine, svm, thesis, masters, honours, facial, expression, recognition, academics, academia, object, detection

Honours Mini-Theses

Car Counting system

Design an implement a “car counting system” that uses videos of highway traffic as input and extracts the number of cars passing the camera per hour.

Topics: image processing, object detection.

Programming: 30%
Research: 70%

System can be extended (by other students or as a separate honours thesis) to handle tracking of vehicles through a network of roads and statistics extraction (such as speed, number of occupants, time taken, damage(?) to vehicle), by maybe having multiple roadside ’stations’ at which the car counting system is installed.

3D game engine based virtual environments

Using an existing open-source 3D game engine, create realistic, virtual, fully 3D interactive environments and evaluate according to modeling abilities, usability and navigability and extensibility. Students are required to create at least 2 maps: a house and a school or computer-science department.

Topics: Modeling, HCI.

Research: 40%
Programming: 60%

Autonomous vehicle driving system: using 1 video camera

Using a form of machine learning (like neural networks or support vector machines) create “DriveIt 1.0″ - an autonomous driving system for vehicles.
Limitations for honours level: Assume a smooth, level driving surface/road with no potholes, adequate lighting conditions, no other vehicles or pedestrians, and maximum road curvature of 90 degrees in 100 meters.

Topics: machine learning, image processing.

Research: 30%
Programming: 70%

Autonomous vehicle driving system: using 2 video cameras

Using a form of machine learning (like neural networks or support vector machines) create “DriveIt 2.0″ - an autonomous driving system for vehicles.
Limitations for honours level: Assume a smooth, level driving surface/road with no potholes, adequate lighting conditions, no other vehicles or pedestrians, and maximum road curvature of 90 degrees in 100 meters.

Topics: machine learning, image processing.

Research: 30%
Programming: 70%

Autonomous vehicle driving system: using a bank of IR sensors

Using a form of machine learning (like neural networks or support vector machines) create “DriveIt 2.5″ - an autonomous driving system for vehicles.
Limitations for honours level: Assume a smooth, level driving surface/road with no potholes, adequate lighting conditions, no other vehicles or pedestrians, and maximum road curvature of 90 degrees in 100 meters.

Topics: machine learning, image processing.

Research: 30%
Programming: 70%