Learning Architectures for Visual Object Recognition

Tuesday, December 16, 2014

 

Infotech Oulu Doctoral Program

Lecturer: Dr. Ross Girshick, University of California, Berkeley, USA & Microsoft Research Cambridge, UK

Date: 16th of December, 2014

9:30-11:30 (lecture room TS101)

Overview of object detection methods
Deformable Part Models in Detail

13:00-16:00 (lecture room TS128)

Deep Convolutional Neural Networks for Recognition and Detection

  • Introduction to convolutional neural networks (basics and history
  • Image classification using deep convolutional networks
  • Object detection using deep convolutional networks
  • Relationship between Deformable Part Models and convolutional networks

Abstract:

This mini-course will cover advances in visual object recognition over the last decade, including the recent wave of results based on deep convolutional networks. Special attention will be placed on the task of object detection. We will cover Deformable Part Models (DPMs) in detail in the first portion of the course. Then we will shift focus to the methods behind the recent advances in recognition performance based on deep convolutional neural networks (ConvNets). The basic structure, methodology, and history of ConvNets will be covered without assuming much background knowledge. Finally, the connection between DPMs and ConvNets will be discussed, giving a unified view of the two recognition methods.

Lecturer’s Biograph:

Ross Girshick finished his Ph.D. in computer vision at The University of Chicago under the supervision of Pedro Felzenszwalb in April 2012. Since then, he's been a postdoctoral fellow working with Jitendra Malik at UC Berkeley.

Ross's main research interests are in computer vision, AI, and machine learning. His work focuses on building models for object detection and recognition that aim to incorporate the "right" biases so that machine learning algorithms can understand image content from moderate- to large-scale datasets.

During Ross's Ph.D., he spent time as a research intern at Microsoft Research Cambridge, UK working on human pose estimation from depth images. He also participated in several first-place entries into the PASCAL VOC object detection challenge, and in 2010 was awarded the PASCAL VOC "lifetime achievement" prize for his work on Deformable Part Models.

More information: Juho Kannala

Add to calendar

Back to events

Last updated: 8.12.2014