V isual I nformation nformation P P rocessing V isual I rocessing Group Group Methods to to Improve Improve Resolution Resolution of of Methods Compressed Video Video Sequences Sequences Based Based on on Compressed Observability Maps Maps Observability Ioannina (Greece), October 2004 By L.D. Alvarez Ioannina, 10/04/2004
L.D. Alvarez -- Ioannina, September 2004 Outline Outline INTRODUCTION • Myself � Introduction • University � A few things about myself. • Research group � A few things about my city and university. RESEARCH � A few things about my research group. • Introduction � Brief introduction. and notation � Staff. • System model � Research areas. • Problem formulation • Optimization � Research procedure • HR simultaneous � Introduction and notation. estimation � System model. • Observability & predictability � Problem formulation. • Results � Optimization procedure. � HR simultaneous estimation. � Observability and predictibility. � Results.
L.D. Alvarez -- Ioannina, September 2004 INTRODUCTION INTRODUCTION
L.D. Alvarez -- Ioannina, September 2004 Myself Myself INTRODUCTION • Myself � Personal information. • University � Luis David Alvarez Corral. • Research group � Granada, Spain, 1977. RESEARCH • Introduction � My background. and notation � Computer Science Engineering degree at the University of • System model Granada in 2001. • Problem formulation � Advanced Studies Diploma in 2003 (Master). • Optimization � Currently pursuing a Ph.D. degree at the department of procedure Computer Science and Artificial Intelligence of the University of • HR simultaneous Granada. Granted by the Spanish Ministry of Education and estimation Science. • Observability & predictability • Results � My research areas. � Enhancement of images and video. � Superresolution problem from compressed observations using the Bayesian framework (problem in compressed video within a Bayesian framework).
L.D. Alvarez -- Ioannina, September 2004 The city city and and the the University University The INTRODUCTION • Myself � A brief history • University � Granada was iberian, roman, jewish, islamic (capital of the former • Research group Nazari Kingdom) and Christian (last city on the Iberian Peninsula to be conquered from the Muslims in 1492 by the Catholic Kings). RESEARCH � The Christian conquest did not rob the city of its splendour as a • Introduction and notation cultural centre, in which sciences and humanities found the best • System model way to develop. • Problem formulation � A historic University • Optimization procedure � The University of Granada was founded in 1531 by Emperor • HR simultaneous Carlos V. estimation • Observability � The University continued the tradition of the Arab University of & predictability Yusuf I. • Results � With 473 years of tradition, the University of Granada has been an exceptional witness to history, as its influence in the city's social and cultural environment grew until it was to become, over a period of almost five centuries, an intellectual and cultural nucleus in Southern Spain in its own right.
L.D. Alvarez -- Ioannina, September 2004 INTRODUCTION � University numbers (year 2002) • Myself � Some 70,000 people are directly linked with the University of • University Granada, among them students, teachers and administrative and • Research group service staff. � Number or degree courses 72 (in 24 university centres and 4 RESEARCH associated) • Introduction and notation � Degree and diploma students 60,000 • System model � Administration and Service Staff 1,578 • Problem formulation � Departments 107 • Optimization � Computers connected to Internet 9,500 procedure • HR simultaneous estimation • Observability & predictability • Results For further information: http://www.ugr.es
L.D. Alvarez -- Ioannina, September 2004 Research group group: VIP : VIP Research INTRODUCTION • Myself • University � The research group “Visual Information Processing” • Research group (VIPG) consists of members of the Departament of Computer Science and Artificial Intelligence and the RESEARCH • Introduction Department of Computer Languages at the and notation University of Granada and the Departament of • System model • Problem Informatics at the University of Jaén. formulation • Optimization procedure � Staff: • HR simultaneous estimation � Professor Nicolás Pérez de la Blanca. • Observability & predictability � Professor Rafael Molina Soriano. • Results � 6 associate lecturers. � 3 assistant lecturers. � 1 associate investigator. � 7 funded Ph. D. student.
L.D. Alvarez -- Ioannina, September 2004 � Some colaborations: INTRODUCTION � Institute of Astrophysics of Andalucía (IAA). • Myself • University � Prof. Brian Ripley (Oxford University, then at Strathclyde • Research group University) on astronomical image restoration. � IMSOR (now IMM) at the Technical University of Denmark RESEARCH on remote sensing classification problems. • Introduction � Prof. A.K. Katsaggelos * (Northwestern University of and notation • System model Evanston, Illinois). � Prof. F. Murtagh * (Queen’s University of Belfast). • Problem formulation � Prof. E. P. Simoncelli * (New York University, New York). • Optimization procedure � Prof. Nikolaos Galatsanos * (Ioannina University, Greece). • HR simultaneous estimation � Prof. E. Trucco (Herriot Watt University, Edinburgh) on 3D modelization and videoconference. • Observability & predictability � Prof. Gustavo Marrero (IUMA, Las Palmas, Spain) on • Results hardware implementation of super-resolution algorithms. � * � On areas such as image representation, image restoration and reconstruction of (compressed/uncompressed) images and sequences.
L.D. Alvarez -- Ioannina, September 2004 � Some research areas: INTRODUCTION � Information retrieval from image databases. Feature • Myself extraction from colour images. • University • Research group � Extraction of 3D information from video sequences for multimedia applications. RESEARCH � Motion analysis. Applications to optical flow estimation and • Introduction and notation motion segmentation. • System model � Tracking algorithms and optical flow. • Problem formulation � Superresolution from compressed and uncompressed • Optimization stills and video sequences. procedure � Visual-statistical approach to multi-scale multiorientation • HR simultaneous estimation representation models. Application to: texture modeling • Observability and synthesis, image restoration, image enhancement, & predictability model-based image coding. • Results � Bayesian image restoration and reconstruction. Applications to Astronomy and Medicine. � Fuzzy image processing. Image classification and pattern recognition.
L.D. Alvarez -- Ioannina, September 2004 INTRODUCTION • Myself For further information: • University • Research group RESEARCH • Introduction and notation http://decsai.ugr.es/vip • System model • Problem formulation • Optimization procedure • HR simultaneous estimation � Members: http://decsai.ugr.es/vip/members • Observability & predictability � Publications : http://decsai.ugr.es/vip/publications (where • Results you can get a copy of all our published work)
L.D. Alvarez -- Ioannina, September 2004 RESEARCH RESEARCH
L.D. Alvarez -- Ioannina, September 2004 Introduction and and notation notation Introduction INTRODUCTION • Myself • University � Objective: Our aim is to estimate high resolution video from a • Research group compressed low resolution image sequence. RESEARCH • Introduction and notation • System model • Problem formulation • Optimization procedure • HR simultaneous estimation • Observability & predictability Compressed • Results Low-resolution Low-resolution LR 128x128 � g k High-Resolution HR 256x256 � f k CLR 128x128 � y k
L.D. Alvarez -- Ioannina, September 2004 System model model System INTRODUCTION • Myself g l • University f l � Objective: To estimate f l • Research group y l RESEARCH l from 1 to the image sequence lenght ( L ) • Introduction and notation • System model • Problem formulation � Relationship between one LR image and the corresponding • Optimization HR image: procedure • HR simultaneous g = AHf estimation l l • Observability & predictability • Results � H � blurring � A � downsampling
Recommend
More recommend
