Skip to main content

Fatemeh Shokrollahi Yancheshmeh: Visual similarity is a tool to automatically search similar images from large image collections

Tampere University
LocationKorkeakoulunkatu 1, Tampere
Tietotalo building, room TB109, Hervanta Campus, and remote connection.
Date2.2.2024 10.00–14.00
Entrance feeFree of charge
Photo: Fatemeh Shokrollahi Yancheshmeh
In her doctoral dissertation, Fatemeh Shokrollahi Yancheshmeh investigates features that match between images of the same category of objects. Visual similarity measure using the local features and their spatial relationships provides a tool to find "visual paths" between any two images, for example, how a bicycle smoothly transforms into a motorbike.

Visual similarity is an important tool to solve many problems in computer and robot vision. For example, humans and robots use familiar landmarks to plan their route and navigate between different locations. Visual similarity is the computatonal tool to recognize a familiar landmark. The doctoral thesis of Fatemeh Shokrollahi Yancheshmeh investigates similarity in the level of object categories.

“Category level similarity is difficult as many objects from the same category look visually very different. For example, pictures of two different persons or cars,” Fatemeh Shokrollahi Yancheshmeh says.

The developed methods can be used to search for similar visual structures in large image databases. One example is finding similar lesions from millions of patient images in a medical image database. Medical doctor needs to annotate only a single example and then the system can provide similar lesions in other images in the database.

Public defence on Friday 2 February

The doctoral dissertation of MSc (Tech) Fatemeh Shokrollahi Yancheshme in the field of information technology titled Visual Similarity for Object Detection and Alignment will be publicly examined at the Faculty of Information Technology and Communication Sciences at Tampere University in room TB109 of the Tietotalo building (address: Korkeakoulunkatu 1, Tampere) at 12:00 on Friday 2nd February 2024.

The Opponents will be Professor Heikki Kälviäinen from LUT University, Finland, and Senior University Lecturer Jorma Laaksonen from Aalto University, Finland. The Custos will be Professor Joni Kämäräinen from the Faculty of Information Technology and Communication Sciences, Tampere University.

Tutustu väitöskirjaan

Seuraa väitöstilaisuutta etäyhteydellä

Kuva: F. Shokrollahi Yancheshmeh