Learning Visual Object Descriptors

D81faf59 9988 47c4 890a ed7c3630b41b.png 500 500

Starting Date: earliest start: (2016-11-01) latest end: (2017-09-30)

Organization: Autonomous Systems Lab

Involved Host(s): Gilitschenski Igor , Cadena Cesar

Abstract: The goal of this project is developing and evaluating a machine learning based full-object descriptor algorithm for object retrieval

Description: Detecting, classifying, and retrieving objects are important tasks in numerous robotics applications involving manipulation, grasping, and navigation. Particularly the latter may benefit from an object-based approach as storing coarse object information may be less prone to lightning and appearance changes. Within this project, we aim at using recent advances in machine learning in order to learn a coarse visual object descriptor that is suitable for object retrieval while at the same time being robust against appearance changes. That is, the aim is to obtain a descriptor that provides more information about the "identity" of the object and not just the object class while at the same time being compact and robust against appearance changes. **What we offer** - Contributing to ongoing research in an exciting emerging field. - Working with one of the largest robotics research teams in the world. - Strong ties with major industrial partners and numerous spin-off companies. - Opportunities to submit your work for publication in case of project success.

Work Packages: - Evaluate existing object detection and classification algorithms. - Build a novel CNN based object descriptor.

Requirements: - Strong C++ and/or Python coding skills. - A good understanding of computer vision, probability, and linear algebra. - Excellent command of machine learning algorithms, particularly (Convolutional) Neural Networks and experience in using typical machine learning frameworks such as TensorFlow or Caffe. - Students from outside of D-MAVT (particularly, from D-INFK, D-ITET, D-PHYS, and D-MATH) are also highly encouraged to apply.

Contact Details: If you are interested, please send your grade transcripts, CV, and a few sentences about your coding background to Igor Gilitschenski (igilitschenski@ethz.ch) and Cesar Cadena (cesarc@ethz.ch).

C++ Python Machine Learning Convolutional Neural Networks (CNNs) TensorFlow Caffe Computer Vision Classification

Labels: Semester project Master Thesis CLS Student Project (MPG ETH CLS)
Topics: Mathematical Sciences Information, Computing and Communication Sciences Behavioural and Cognitive Sciences

© 2017, Copyright Max-Planck-Gesellschaft - Imprint