Course Topic What is Computational Photography ? Study the basics - PDF document

Course Topic • What is “ Computational Photography ”? • Study the basics of computation and its impact on the entire Computational Photography workflow of photography, from – capturing, – manipulating, and – collaborating on, and CS 4475/6475 – sharing photographs. Maria Hybinette • Focus is on the computational pipeline not how to use the camera – but we will do some usage too… 1 Maria Hybinette 2 Maria Hybinette Where are the Algorithms? Who am I • Photography : Started out with interest in the • Using ‘ computations ’ to create an image. ‘artistic’ form of photography and expanded studying • Using ‘ computations ’ to extend the hardware the computational aspect in 2005-2006 – Felt understanding the science helps understanding the ‘art’ – On camera at a deeper level – Off-camera – Sold photographs a number of photographs through Getty Images (both Stock and Creative photographs, their categories). – More of a hobby now but occasional use flickr • https://www.flickr.com/photos/cooriander/ • Academics : Emory University • Graduated from G-Tech in 1995, and 2000 • Teaches at the University of Georgia 3 Maria Hybinette 4 Maria Hybinette Today Learning Tools & Evaluated • Overview of the materials that we will cover in the • Again it is about the computations in the course photography • Expected background – Learn via hands-on experimentation • and what to except from the course • Frequent Homework • Tomorrow : Assignments (almost weekly) – Deeper Dive into computational – Coding mostly photography • 3 mini projects through the – 2 examples putting computational summer photography in context – Free form • Dual photography • Quizzes 3 • Panorama – Python/OpenCV discussion setup. 5 Maria Hybinette 6 Maria Hybinette

Communication Links Expected Background & Platforms • Math Knowledge • Piazza – Linear Algebra (matrices, vectors) • T Square – Calculus (differential and integration) • Schedule Page – Probability (sampling, and distributions functions) • Computation – Python (easy to pick up) – OpenCV (CV = computer vision) • Hardware – Laptop – Fancy Camera (not required), but you will need access to some image generating device • Control: Exposure, Color, Focus, and Lenses (Length) h"p://cobweb.cs.uga.edu/~maria/classes/2016-CompPhoto/schedule.html – Images will be provided in most cases. 7 Maria Hybinette 8 Maria Hybinette Wk1: Introduction Theme: Computational Photography • What is computational • Buildings the computational technology and tools photography? how to figure out : • Computational Photography • How light is used to produce images and as a discipline photographs. • 2 practical Examples of – Rays to Images (pixels) Computational Photography providing context. – Images to Sharing • Assignment – Sharing pictures both the good and the bad – Getting set up. Python, OpenCV h"p://graphics.cs.cmu.edu/?page_id=16 9 Maria Hybinette 10 Maria Hybinette Wk 4: Camera, Optics & Sensors Wk 2-3: Image Processing and Analysis • Digital Image Representation • Digital Image • Pixel/Point processes for Representation Pixel/Point Processes for Images images • Smoothing and filtering • Smoothing and Filtering methods for Images methods for Images • Extracting features from • Extracting Features from Images Images • Assignments : Experiments • Assignments : with Image Filtering, Features Experiments with Image Detection Filtering, Features Detection 11 Maria Hybinette 12 Maria Hybinette

Wk 5: Image Blending/Merging Wk 6-X: More Advanced Concepts • Sampling and • More advanced concepts Frequencies • Feature Detection • Image Blending • Panoramas • Homework: Exercise on • HDR Image Blending • Image Editing. 13 Maria Hybinette 14 Maria Hybinette Limitations of Photography Computational Photography Combines • Computing • Blur, camera shake, noise, damage • Digital Sensors • Modern Optics • Actuators – control aspects of photography • Smart Lights • To “escape”the limitations of traditional film cameras h"p://www.cs.unc.edu/~lazebnik/research/fall08/ 15 Maria Hybinette 16 Maria Hybinette Limitations of Photography Limitations of Photography • Limited Resolution • Bad color / monochromatic h"p://learningdslr.com/2010/10/colour-temperature-white-balance/ h"p://en.wikipedia.org/wiki/Black-and-white 17 Maria Hybinette 18 Maria Hybinette

Limitations of Photography Limitations of Photography • Unwanted objects • Limited dynamic range h"p://www.glosonblog.com/adobe-photoshop-how-to-remove- unwanted-items-from-photos/ h"p://cpn.canon-europe.com/content/educaIon/technical/ h"p://en.wikipedia.org/wiki/High_dynamic_range_imaging photoshop_cs3_2.do 19 Maria Hybinette 20 Maria Hybinette Limitations of Photography Elements of Computational Photography • Single depth of focus • Computation can be embedded in all aspects of these elements to support photography rays to pixels to sharing. • Enhance the photographic process – Rays to pixels h"p://www.janrik.net/insects/ExtendedDOF/LepSocNewsFinal/EDOF_NewsLepSoc_2005summer.htm Source: Irfan Essa 21 Maria Hybinette 22 Maria Hybinette Sources & Inspiration Tomorrow • Deeper dive into computational Contributors of Course Material: photography • Marc Levoy (Stanford)– taught computational photography since 2002: • 2 examples putting computational – A leader in the field : Frankecamera photography in context • Irfan Essa & Frank Dellaert (Georgia Tech) – Dual photography – Also early adopters – Field Trip, Guest Lecturer (HDR) – Panorama • Frédo Durand (MIT) • Computational photography and how • Jack Tumblin (Northwestern) it relates to other disciplines. • Wikipedia • http://www.all-art.org/ • Python/OpenCV discussion setup. history658_photography1.html • “Photography”, London, Stone, Upton 23 Maria Hybinette 24 Maria Hybinette

Computational Photography Study the basics of the impact of computation on the entire workflow of photography, from how images are captured, manipulated and collaborated on, and shared. 25 Maria Hybinette