Lunds Tekniska Högskola

Denna sida på svenska This page in English

Extra CS MSc Presentations June 15!


Two MSc theses to be presented on June 15, 2018.

June 15 is an extra day for master thesis presentations in Computer Science at Lund University, Faculty of Engineering. Two MSc theses will be presented.

The presentations will take place in the E-house, room E:4130. A preliminary schedule follows.

Note to potential opponents: Register as opponent to the presentation of your choice by sending an email to the examiner for that presentation ( Do not forget to specify the presentation you register for! Note that the number of opponents may be limited (often to two), so you might be forced to choose another presentation if you register too late. Registrations are individual, just as the oppositions are! More instructions are found on this page.


E:4130 (LUCAS)


PRESENTERSLinus Röman, Simon Finnman
TITLEAlgorithmic Approach to Error Correction in Map Datasets using Conflation Techniques
EXAMINERFlavius Gruian
SUPERVISORKrzysztof Kuchcinski (LTH), Thomas Hermansson (ÅF - Digital Solutions AB)

OpenStreetMap is a crowd sourced, free and open sourced data-set that contains geographical data. As any other data source OpenStreetMap contains a variety of errors, topological, geo-spatial and semantic to mention a few. This thesis focuses on using algorithms to detect and flag these errors. In particular algorithms built on the concept of conflation, eg. using another data-set in order to compare. The first step is to establish a matching between the two data-sets, which is done with an accuracy of above $94%$ in our test areas. After this match has been established differences between the data-sets can be found. These differences can be used to flag errors which can be forwarded for manual correction. We have looked at attributes such as names, where name dissimilarity can been used to differentiate different types of errors from each other, as well as investigated the correctness of speed limits in the OpenStreetMap.


PRESENTERSEmir Husic, Aleksandar Simeunovic
TITLEApplying Object Recognition To Reduce False Alarms Triggered By Motion Detection In Video Surveillance
EXAMINERFlavius Gruian

Using motion detection in surveillance cameras is a decent way of detecting actions in environments. However, it is incapable of determining the causing source, such as animals, flying objects, or humans. This incapability tends to trigger alarms where more often then not, a human is not present.

In this report, we study the effect of adding another evaluation layer before triggering an alarm - an object detection layer identifying humans explicitly. Video alarms triggered by motion detection, and simultaneously containing human properties are weighed to be more likely of a real alarm.

We present methods, and choices of data used while applying object detection which filters up to 85% of false alarms without losing true ones.

The benefit of our approach is the reduction of human hours spent on evaluation of false alarms. It is also possible to train the detector to specific environments increasing the accuracy by using neural networks.