Topics within our research lines for B. Sc. thesis (TFG) and M. Sc. thesis (TFM):
TFG: Network probe over ALTO
B.Sc thesis (TFG)
OBJETIVE:
ALTO is a standard solution defined in IETF [1] for exposing network topology maps with associated metrics between points on a network identified by IP address prefixes. One of the advances proposed on the ALTO specification is the analysis of events in the network (link drops, packet losses, delays, etc.) using Machine Learning and Artificial Intelligence techniques to predict behavior and make shipping decisions. of information in the network [2].
This TFG proposes to increase ALTO with the addition of network metrics collected in the different events created to implement complex data analysis that allow optimization in network management, for example, in sending content through CDNs.
To do this, it is proposed to add new network metrics collected from the different events created and use network tools to calculate this information and transmit it to a central server. A comparison of network tools will be carried out and statistics of the data collected will be provided in a results viewer such as Grafana.
As future work, other Machine Learning and Artificial Intelligence techniques could be explored to further improve ALTO’s analysis capabilities.
[1] https://www.rfc-editor.org/rfc/rfc7285.html
[2] https://web.archive.org/web/20220827172558id_/https:/dl.acm.org/doi/pdf/10.1145/3538401.3546602
REQUIREMENTS:
Python, Linux knowledge (terminal and script execution), network knowledge (core, protocols in general).
CONTACT PERSON:
Alberto del Río Ponce – arp@gatv.ssr.upm.es
TFG / TFM: Image scaling techniques in audiovisual content
B.Sc thesis (TFG), M.Sc. thesis (TFM)
OBJECTIVE:
Image scaling is an essential process in image processing. This process consists of a change in the size of the image. On many occasions, the resolution of the image must be enlarged or reduced, which implies a change in the number of pixels.
In this work, different traditional scaling algorithms will be tested, such as bilinear interpolation, bicubic interpolation and nearest neighbor interpolation, among others. The possibilities offered by new methods based on conventional neural networks will also be explored. Finally, quality metrics (objective and subjective) will be used to compare the results obtained after applying the different types of image scaling.
REQUIREMENTS:
Knowledge of image processing and neural networks. Programming notions in Python.
CONTACT PERSON:
Álvaro Llorente – alg@gatv.ssr.upm.es
TFG: Development with TSDuck
B.Sc thesis (TFG)
OBJETIVE:
Development of a tool that allows analyzing and modifying both the contents and the service information tables in an MPEG Transport Stream. For this, the TSDuck tool will be used, an Open Source tool used in digital television systems, written in C++ although it has a large collection of plugins that can be used through the command line. The tool will be integrated with professional equipment that this research group has, both in reception and transmission of multimedia content through DTT.
REQUIREMENTS:
Notions of programming in Python for the integration of the tool.
CONTACT PERSON:
Álvaro Llorente – alg@gatv.ssr.upm.es