Details

Analiza vpliva zamika koncepta na uspešnost metod za detekcijo površinskih anomalij
ID Pahor, Jure (Author), ID Skočaj, Danijel (Mentor) More about this mentor... This link opens in a new window

.pdfPDF - Presentation file, Download (10,51 MB)
MD5: 9BDF6E163DE25384D71CCBFBEBBEBCC2

Abstract
Diplomska naloga obravnava vpliv časovnih sprememb v okolju in na samih predmetih na metode nenadzorovane detekcije površinskih anomalij. Osredotoča se na tako imenovani zamik koncepta (ang. concept drift), ki pogosto otežuje odkrivanje napak v različnih okoliščinah. V sklopu naloge so podrobneje preučeni vplivi sprememb na slikovnem gradivu, in sicer na področjih jakosti osvetlitve, barvne temperature osvetlitve, refleksivnosti materiala, stopnje nehomogenosti normalnosti izgleda oziroma šuma ter različno izraženih anomalij na predmetih. Cilj je ugotoviti, kako tovrstne spremembe vplivajo na učinkovitost metod nenadzorovane detekcije površinskih anomalij. V ta namen smo s pomočjo orodja Blender sintetično generirali podatkovno množico slik, ki zajema naštete vidike sprememb in omogoča popoln nadzor nad vsemi parametri. Želimo preveriti, ali se vnaprej naučeni modeli za detekcijo anomalij uspešno soočajo z manjšimi, a postopnimi spremembami v okolju. Da bi to dosegli, najprej ovrednotimo velikost posameznih sprememb, nato pa dobljene vrednosti primerjamo z rezultati izbranih modelov za detekcijo anomalij in iščemo morebitno korelacijo med njimi.

Language:Slovenian
Keywords:detekcija anomalij, računalniški vid, globoke nevronske mreže, površinske anomalije, zamik koncepta
Work type:Bachelor thesis/paper
Typology:2.11 - Undergraduate Thesis
Organization:FRI - Faculty of Computer and Information Science
Year:2025
PID:20.500.12556/RUL-168182 This link opens in a new window
COBISS.SI-ID:231418627 This link opens in a new window
Publication date in RUL:01.04.2025
Views:72
Downloads:18
Metadata:XML DC-XML DC-RDF
:
PAHOR, Jure, 2025, Analiza vpliva zamika koncepta na uspešnost metod za detekcijo površinskih anomalij [online]. Bachelor’s thesis. [Accessed 8 April 2025]. Retrieved from: https://repozitorij.uni-lj.si/IzpisGradiva.php?lang=eng&id=168182
Copy citation
Share:Bookmark and Share

Secondary language

Language:English
Title:Analysis of the impact of concept drift on the performance of surface anomaly detection methods
Abstract:
This thesis examines the impact of temporal changes in the environment and on the objects themselves on methods for unsupervised surface anomaly detection. It focuses on the so-called concept drift, which often complicates fault detection in various contexts. The study delves into how changes in image data—including illumination intensity, color temperature, material reflectivity, the degree of non-uniformity of normal appearance (i.e., noise), and variously pronounced anomalies on objects—affect the performance of these methods. To investigate this, a synthetic image dataset was generated using Blender, which allows complete control over all parameters related to the aforementioned changes. The aim is to determine whether pre-trained anomaly detection models can effectively handle minor yet gradual environmental shifts. To do so, the magnitude of each change is first evaluated, and these values are then compared with the results produced by selected anomaly detection models to explore any potential correlations.

Keywords:anomaly detection, computer vision, deep neural networks, surface anomalies, concept drift

Similar documents

Similar works from RUL:No similar works found
Similar works from other Slovenian collections:No similar works found

Back