KEYWORD |
Near duplicate image detection
Thesis in external company
Reference persons FABRIZIO LAMBERTI
Research Groups GR-09 - GRAphics and INtelligent Systems - GRAINS
Description Detecting images that are near duplicates (that is slightly altered versions of the same scene) has many important applications (forgery detection, web image retrieval, etc.) This thesis will investigate and compare different approaches that include global and local scale analysis, using both establish techniques (SIFT, SURF, etc.) and neural descriptors from deep convolutional neural networks. Another important aspect is how to efficiently represent, store and navigate large databases that may contain near duplicates. Skills possessed or to acquire: programming skills (openCV, C++ or Python, Keras/Tensorflow or other deep learning framework); analytical skills, detail-oriented attitutes. This thesis could be carried out in collaboration with a company.
See also http://grains.polito.it/work.php
Deadline 13/02/2019
PROPONI LA TUA CANDIDATURA