KEYWORD |
DAUIN - GR-04 - DATABASE AND DATA MINING GROUP - DBDM
Color grading style classification through Vectorscopes and CNNs
keywords COMPUTER VISION, SUPERVISED DEEP LEARNING, COLOR G
Reference persons TANIA CERQUITELLI
External reference persons VACCHETTI BARTOLOMEO (https://www.dauin.polito.it/it/personale/scheda/(nominativo)/bartolomeo.vacchetti)
Research Groups DAUIN - GR-04 - DATABASE AND DATA MINING GROUP - DBDM
Thesis type DATA ANALYSIS & MODELING
Description Color has a deep impact on the viewer's perception of a movie. This project aims to characterize the different color styles of films in relation to their genre using vectorscopes. A vectorscope is a commonly used tool for color grading and color correction. The student will:
++ create a vectorscope dataset
++ train a neural network to test the methodology
++ analyze and interpret the results
See also https://smartdata.polito.it/creativity-injection-into-ai-powered-multimedia-storyboards/
Required skills Python programming, image processing, Machine Learning techniques
Deadline 31/07/2023
PROPONI LA TUA CANDIDATURA