Alireza Tavakoli Targhi's Homepage
Alireza Tavakoli Targhi's Homepage
Projects:

MOBVIS: European Project
2007-2008
The main objective of the project is to explore and exploit the concept of attentive interfaces for the design and implementation of "mobile vision technology".

MUSCLE: The EU Network of Excellence
2004-2007
This network is made up of 42 member laboratories with in European countries, collaborating on various approaches to extracting semantics from image and video data.
Collaborators:
Jan-Mark Geusebroek, Nicu Sebe,
Sven Dickinson, Alex Levinshtein
• Computer Vision and Machine Learning, university of Bon, Germany

Research:

This project is part of a European Project, Mobvis. The goal of this master project is to extract text information from an image which is taken from plate streets via a Mobile phone’s camera. The scenario for this project is that a tourist in an urban environment is equipped with a mobile phone’s camera in order to be able to take a picture from any text information such as street’s name, next is our software which is the goal of this project, segmenting the text region from the image and by using an OCR software (Optical character recognition) attempting to recognize the text. This text can be used as a query for search over several databases to get more information about the place or object for instance spot the location of an object or person on the map by the usage of GPA, Show the location of the person on the MAP with help of GPA, explore for more information through the internet, and search for the general information of the designated place, etc.

This is a collaboration within Network of Excellence with other European universities. Image search and retrieval is a challenge in computer vision besides being a demand for the industry. The goal of this project is search though another image to detect a specific object. For example if you search "cat" on Google any file which is named "cat" will appear where as many of them is not related to cats. In this project, instead of looking on name of the image file, we search through image and analyze the inner part of the images to detect the cat. Therefore we are going to define as cat by the set of features such as "texture", "shape", "Colour", etc. We would use a learning method to study the appearance of the animal class which would be described by these set of features. Accordingly, for Animal recognition we will work with "feature extraction" and different learning method.

The recognition of materials from their visual texture has many applications, for instance it facilitates image retrieval and object recognition. The aim of this project is to use information of material of objects for a service robot. The scenario for this project is that when we ask a service robot “pickup the bread from the breakfast table”, the robot be able to distinguish between "sponges", "brown-bread", "cotton clothed", “corduroy cloths ”, or “aluminum box” from their visual texture. Therefore the master student should work with computer vision and robotic for this project
Postal addressNADA/CVAP
KTH
S-100 44 Stockholm
Sweden
Office addressRoom 6.12
Teknikringen 14
Stockholm
Phone+46 (0)8 790 6725
e-mailattkth.se