The ML summer school is slightly theoretical in nature but sufficient practical exercises would be covered to enable a better understanding of the theoretical track. Primary theme for this year’s summer school is chosen to be Deep Learning, seeing the recent trends of its rise. Experts in the field will deliver talks to share their views and works with the attendees and you are more than welcome to interact and discuss ideas with the speakers. It will also be a good platform to bounce ideas amongst other like-minded enthusiasts. The primary focus of the summer school this time would be on the more recent advancements in the field of Deep Learning.
The summer school curriculum roughly consists of a series of lectures and demo/labs sessions designed to work in tandem to help you make the most out of the program. Experts in the field will deliver talks to share their views and works with the attendees and you are more than welcome to interact and discuss ideas with the speakers. It will also be a good platform to bounce ideas amongst other like-minded enthusiasts. This time we aim to have a special focus on the recent advances in the area.
Deep learning has resulted in the best solutions for many challenging computer vision problems in recent years. The course on Deep Learning at IIIT Hyderabad aims to keep the pace with the rapid growth in this field, and expose the advances to working professionals and researchers. The course will focus on foundations, recent advances with special emphasis to running on limited memory platforms and the practical aspects of using deep learning for a variety of computer vision problems.
Many problems in document image analysis are being formulated as maximum a posteriori (MAP) estimation in a Markov/conditional random field (MRF/CRF) setting. Examples include restoration, binarization, segmentation,and recognition. This class of formulations has resulted in the development of elegant algorithms for many challenging problems in this area. Often the output of these inferencing algorithms is a structured description (e.g., a string of characters, an array of labelled pixels, a tree/graph description of a document image). In this tutorial, we would like to connect the two well established areas (i.e., document image analysis, structured prediction by MAP estimation), and demonstrate how systematic development of robust algorithms can be enabled.
Scalability of a given solution is an important consideration towards enabling retrieval and recognition over large collections of document images. However, the definitions of scalability are fast changing with the emergence of huge datasets and digital libraries, as well as the advent of new computing paradigms. In this tutorial, we shall cover three approaches towards building scalable document image retrieval and recognition systems:
- Recognition-free retrieval using bag-of-visual-words
- Recognition of word-images using indexing schemes
- Large-scale testing/deployment using cloud computing
This tutorial shall include a parallel hands-on practical session, where the attendees would have the opportunity to practice the methods described in the tutorial. A dataset, along with the necessary code libraries, will be provided to the audience. Multiple solution stacks shall be deployed and evaluated by the various groups/individuals, with a scalable retrieval system being built by the end of the tutorial.
Workshop on Computer Vision 2008, WCV '08 was organized under the Indo-Israeli MoU on research collaboration managed by DST. The workshop was held at IIIT Hyderabad from February 04, 2008 to February 05, 2008. Top researchers from India and Israel participated in the two day event. The respose to the workshop was overwhelming.
7th Asian Conference on Computer Vision 2006, ACCV '06 was held from January 13, 2006 to January 16, 2006 at Hyderabad. The conference was organized by IIIT Hyderabad. The past conferences in this biennial series were held in Korea (ACCV'04), Australia (ACCV'02), Taiwan (ACCV'00), Hong Kong (ACCV'98), Singapore (ACCV'95), and Japan (ACCV'93).
Course at CVPR 2009, The GPUs have emerged as a useful computing co-processor that is readily available and economical. The latest commodity GPUs are rated for a peak performance of around 1 TFLOPs at the cost of $400 or so. The recent demand for high performance techniques has led to the adaptation of GPUs for various computer vision algorithms. Many computer vision algorithms are well-suited for processing on the GPU due to the match of the data-parallel computations to many operations on images. Recent advances such as CUDA and the OpenCL standard have the potential to accelerate the use of GPUs in many areas for more general purpose computing, including Computer vision. This course aims to familiarize computer vision researchers with the emerging and exciting area of fast computer vision algorithms on the GPU. It will give an introduction to the programming of the current state-of-the-art hardware to enable participants to employ the unique capabilities of GPUs.