Aus Aifbportal
Wechseln zu:Navigation, Suche

Analyse the Impact of Text Representation Methods for Iconclass Label Recommendation

Informationen zur Arbeit

Abschlussarbeitstyp: Bachelor
Betreuer: Harald SackEtienne Posthumus
Forschungsgruppe: Information Service Engineering
Partner: FIZ Karlsruhe
Archivierungsnummer: 4903
Abschlussarbeitsstatus: Offen
Beginn: 20. Mai 2022
Abgabe: unbekannt

Weitere Informationen

ICONCLASS [1] is the de facto global standard for the subject classification of cultural heritage content. It consists of alphanumeric “notations” which document the subjects in images in a language independent way. Each one of these notations has a set of textual labels to describe iconographic subjects in natural language. In your thesis, you will have the chance to evaluate different representation methods for text, such as word embeddings [2], in the task of retrieving correct ICONCLASS codes from textual queries. You will gain valuable experience in NLP and Information Retrieval techniques as well as provide fruitful insights on how to improve the current ICONCLASS text search system. [1] [2]

Which prerequisites should you have? • Good programming skills in Python • Knowledge of machine learning and NLP libraries (e.g., scikit-learn, NLTK)

Download: Download (pdf)