W&M Featured Events
This calendar presented by
William & Mary
[PAST EVENT] Han Li - Oral Exam
February 24, 2015
9am - 11am
Abstract:
With the popularity of digital camera and photo sharing websites, millions of photographs are uploaded to the internet everyday. Large online photo collections provide a rich visual record of our world. Due to the uncontrolled manner in which this data is collected, extracting relevant information from such massive visual data sets poses new research challenges.
Camera response functions characterize how different camera models translate observed irradiance to pixel intensity values, and it is essential to undo its effect when comparing and processing photographs. Most existing methods for estimating the response function of a particular camera require careful calibration under controlled conditions and are not suited for large photo collections. This thesis focuses on estimating camera response functions for each photograph's camera in large photo collections.
We present an example-based approach for radiometrically linearizing photographs that takes as input a radiometrically linear exemplar image, and a target regular uncalibrated image of the same scene, possibly from a different viewpoint and/or under different lighting.
The output of our method is a radiometrically linearized version of the target image. Modelling the change in the appearance of a small image patch seen from a different viewpoint and/or under different lighting as a linear 1D subspace, allows us to recast radiometric transfer in a similar form as classical radiometric calibration from exposure stacks. The resulting radiometric transfer method is lightweight and easy to implement, and we demonstrate the accuracy and validity of our method on a variety of scenes.
Bio:
Han Li has been working on her Ph.D. in Computer Science at the College of William & Mary since Fall 2011. She is working with Dr. Pieter Peers, and her research interests include computer graphics and computer vision. Han Li received her B.S. from Wuhan University (China, 2011).
With the popularity of digital camera and photo sharing websites, millions of photographs are uploaded to the internet everyday. Large online photo collections provide a rich visual record of our world. Due to the uncontrolled manner in which this data is collected, extracting relevant information from such massive visual data sets poses new research challenges.
Camera response functions characterize how different camera models translate observed irradiance to pixel intensity values, and it is essential to undo its effect when comparing and processing photographs. Most existing methods for estimating the response function of a particular camera require careful calibration under controlled conditions and are not suited for large photo collections. This thesis focuses on estimating camera response functions for each photograph's camera in large photo collections.
We present an example-based approach for radiometrically linearizing photographs that takes as input a radiometrically linear exemplar image, and a target regular uncalibrated image of the same scene, possibly from a different viewpoint and/or under different lighting.
The output of our method is a radiometrically linearized version of the target image. Modelling the change in the appearance of a small image patch seen from a different viewpoint and/or under different lighting as a linear 1D subspace, allows us to recast radiometric transfer in a similar form as classical radiometric calibration from exposure stacks. The resulting radiometric transfer method is lightweight and easy to implement, and we demonstrate the accuracy and validity of our method on a variety of scenes.
Bio:
Han Li has been working on her Ph.D. in Computer Science at the College of William & Mary since Fall 2011. She is working with Dr. Pieter Peers, and her research interests include computer graphics and computer vision. Han Li received her B.S. from Wuhan University (China, 2011).
Contact
[[vlthompsondopp]]