HTL Perg DSpace Server
DSpace ist ein digitaler Dienst der Informationen und Wissen sammelt, archiviert und verteilt. Digitale Wissensdatenbanken sind ein wichtiges Werkzeug um Organisationen in der Kommunikation und Verteilung, sowie Weiterentwicklung von Informationen zu unterstützen.
Communities in DSpace
Select a community to browse its collections.
- Dokumente der HTL-Perg
Recent Submissions
ProcessAeye - Echtzeit Kamera Inpainting
(2024) Bauer, Cedric; Czepl, Stefan
The core idea of ProcessAeye is to restore missing or damaged parts of an image in real-time. The goal is to develop an application that reads images from a camera and restores them based on the missing areas, which are defined by the user.
To achieve this goal, three main components are required:
The NVIDIA Jetson Nano is a small, powerful computer designed specifically for AI applications, enabling real-time processing and restoration of images through machine learning models.
Through a Deep Learning model or a classical image processing algorithm, the images are reconstructed. Various models and classical algorithms have been compared in terms of the time required per restoration run and the quality of the results, with the best approaches being used for real-time restoration.
The reconstructed images are displayed in a graphical user interface alongside the original image. In the original image, the user can mark the missing parts of the image using a drawing module. To clearly demonstrate the difference between classical algorithms and deep learning models, users can choose from six different approaches, including three classical algorithms and three machine learning models.
The result is a desktop application optimized for the NVIDIA Jetson Nano which offers six approaches to image restoration. Because the user can draw the mask themselves, the application is ideally suited to illustrate the difference between classical algorithms and deep learning models.
OCR-PostProcessing
(2024) Haiden, Simon; Heindl, Andreas; Hofer, Lukas; Primetshofer, Julian
The OCR post-processing project involves the improvement of text generated by text recognition.
The goal is to achieve the best possible improvements using various approaches. For
implementation four components are required:
The backend, which tries to enhance the given texts using three different approaches, including
two dictionary methods and one AI-based improvement. Additionally, the backend calculates
statistics to assess the performance of each system.
Improvements are tested and visualized through the frontend, which allows for the uploading
of multiple files or entire datasets at once to be improved with all desired correction systems.
All changes are displayed after the improvement process is completed and the backend-calculated
statistics are visualized.
An API connects the user interface and the backend. In Python, using Flask, several endpoints
were defined to facilitate the exchange of information between the frontend and the backend.
Furthermore, a test pipeline allows for the improvement to be used and tested without a
frontend. This pipeline can process a predefined folder structure, correcting all files contained
and comparing them with the ground truth.
The result is a comprehensive application composed of Python modules, JavaScript code, and
PowerShell scripts.
Share:D
(2024) Grillenberger, Noah; Hintermayr, Michael; Lanzinger, Paul; Oprea, Paul
Share:D is a web tool that makes it easier for employees to book shared desks.
These shared desks are located at all Porsche Informatik sites and can be booked by all em-
ployees. These are often used by employees who work in the home office model, but also by
those who move between locations of the Porsche Informatik and therefore do not have a
fixed desk at the other office.
The employees of Porsche Informatik also have a lot of appointments therefore it is easy to
lose track, which is why our web tool also offers an overview of upcoming events. You can also
create a new event here.
The tool is also intended to provide managers with statistical analyses of the bookings made,
such as the utilization of the available shared desks. Among other things, this should help to
determine the amount of space required.
The tool also offers the option of displaying and adding Outlook events.
TRAAKY
(2024) Jandl, Niclas; Kühberger, Clemens; Mayrhofer, Lukas
This thesis describes the development of a web application that is part of a larger collaborative project with FH-Sankt Pölten. The project aims to provide businesses with a cost-effective solution for asset tracking. Rather than engaging costly consultants, users can complete a form on the website and receive insights into three implementations by other companies with similar requirements.
The implementation includes a web interface that is hosted internally at FH. The interface was developed using HTML, CSS, and Bootstrap. Users are required to fill out a form, and the data is processed by a K-Means algorithm that is implemented in PHP and stored in a MYSQL database. The algorithm identifies the three most similar solutions from a set of use cases, and the results are displayed on the website. Furthermore, the website can provide recommendations using the Large Language Model ChatGPT.
The result is a website that offers businesses without asset tracking systems an introduction to suitable localization methods. The website operates smoothly on FH St. Pölten's internal systems. To ensure high accuracy and understanding, users receive assistance from a supervisor when filling out the form. The PHP Slim Framework API facilitates seamless communication between the user interface and backend services. This project offers a solution that is both effective and cost-efficient for businesses of all sizes to find the most appropriate localization method.
LogSense
(2024) Borbely, Philipp; Ettlinger, Sarah; Jilek, Thomas; Stadlbauer, Emily
The basic idea of LogSense is to merge the screen time management of smartphones
and some functionalities from the Windows operating system. As an additional feature,
unexpected behaviour such as high resource consumption of particular processes can be
detected and reported. Another key feature is the long-term storage of the gathered data
so that it can be viewed and analysed retrospectively. To achieve this goal, 4 main components
are required:
The agent runs as a background process on the client computer, collects data about the
hardware components of the PC and sends it to the server for further analysis.
The recorded information is then persisted in the database for further analysis. As this
involves large amounts of time series data, a powerful database is required. The database
most suitable for this is TimescaleDB, since it was developed precisely for this purpose.
The measured data is analysed on the server through the use of machine learning algorithms
and statistical methods. This includes the identification of anomalies, trending
events and the prediction of free storage space.
The graphical user interface displays the recorded data and analysis results in form of
diagrams and statistics. In addition, the user can interact with the system via the user
interface and define customized warnings for a device.
The result is a system consisting of the components listed above, which can be used to
monitor the resource usage of computers as well as the runtime and resource consumption
of individual processes. The collected data is analysed and evaluated in order to be
displayed in the form of statistics, events and anomalies.