Image Processing: Motivation Rendering from Images. Related Work. Overview. Image Morphing Examples. Overview. View and Image Morphing CS334

Similar documents
Image Based Rendering

Image Transfer Methods. Satya Prakash Mallick Jan 28 th, 2003

Image Base Rendering: An Introduction

Stereo pairs from linear morphing

Image-Based Modeling and Rendering

Image-Based Rendering. Johns Hopkins Department of Computer Science Course : Rendering Techniques, Professor: Jonathan Cohen

Image Morphing. The user is responsible for defining correspondences between features Very popular technique. since Michael Jackson s clips

Image Morphing. Application: Movie Special Effects. Application: Registration /Alignment. Image Cross-Dissolve

Image-based modeling (IBM) and image-based rendering (IBR)

CS 4620 Midterm, March 21, 2017

Image-Based Rendering. Johns Hopkins Department of Computer Science Course : Rendering Techniques, Professor: Jonathan Cohen

Announcements. Mosaics. How to do it? Image Mosaics

Image-Based Modeling and Rendering. Image-Based Modeling and Rendering. Final projects IBMR. What we have learnt so far. What IBMR is about

Image-Based Rendering. Image-Based Rendering

Chapter 4. Chapter 4. Computer Graphics 2006/2007 Chapter 4. Introduction to 3D 1

Automatic 2D-to-3D Video Conversion Techniques for 3DTV

But First: Multi-View Projective Geometry

Model-Based Stereo. Chapter Motivation. The modeling system described in Chapter 5 allows the user to create a basic model of a

Photo Tourism: Exploring Photo Collections in 3D

Digitization of 3D Objects for Virtual Museum

Rendering. Converting a 3D scene to a 2D image. Camera. Light. Rendering. View Plane

Soft Shadow Maps. Soft Shadow Maps for Linear Lights

A Review of Image- based Rendering Techniques Nisha 1, Vijaya Goel 2 1 Department of computer science, University of Delhi, Delhi, India

Image warping and stitching

Image-Based Rendering using Image-Warping Motivation and Background

Image warping and stitching

Graphics for VEs. Ruth Aylett

Expression Morphing Between Different Orientations

CSE 167: Lecture 11: Textures 2. Jürgen P. Schulze, Ph.D. University of California, San Diego Fall Quarter 2011

A Statistical Consistency Check for the Space Carving Algorithm.

The Light Field and Image-Based Rendering

Stereo vision. Many slides adapted from Steve Seitz

CS 354R: Computer Game Technology

Image-Based Rendering

Volume Shadows Tutorial Nuclear / the Lab

Graphics for VEs. Ruth Aylett

CSE528 Computer Graphics: Theory, Algorithms, and Applications

View Interpolation for Dynamic Scenes

Introduction to Computer Graphics with WebGL

Models and Architectures. Ed Angel Professor of Computer Science, Electrical and Computer Engineering, and Media Arts University of New Mexico

Topics and things to know about them:

Models and Architectures

Point-Based Rendering

Computer Graphics. Lecture 9 Environment mapping, Mirroring

CHAPTER 3 DISPARITY AND DEPTH MAP COMPUTATION

Lecture 7: Image Morphing. Idea #2: Align, then cross-disolve. Dog Averaging. Averaging vectors. Idea #1: Cross-Dissolving / Cross-fading

Image Morphing. CSC320: Introduction to Visual Computing Michael Guerzhoy. Many slides borrowed from Derek Hoeim, Alexei Efros

Character Modeling IAT 343 Lab 6. Lanz Singbeil

Image-Based Rendering and Light Fields

CS 498 VR. Lecture 18-4/4/18. go.illinois.edu/vrlect18

CS4670: Computer Vision

Hybrid Rendering for Collaborative, Immersive Virtual Environments

Illumination Models & Shading

Modeling, Combining, and Rendering Dynamic Real-World Events From Image Sequences

Spline Morphing. CG software practical course in the IWR. Carl Friedrich Bolz. Carl Friedrich Bolz

Announcements. Mosaics. Image Mosaics. How to do it? Basic Procedure Take a sequence of images from the same position =

Shape Modeling with Point-Sampled Geometry

Finally: Motion and tracking. Motion 4/20/2011. CS 376 Lecture 24 Motion 1. Video. Uses of motion. Motion parallax. Motion field

Jingyi Yu CISC 849. Department of Computer and Information Science

Direct Rendering of Trimmed NURBS Surfaces

CHETTINAD COLLEGE OF ENGINEERING & TECHNOLOGY CS2401 COMPUTER GRAPHICS QUESTION BANK

3D Rasterization II COS 426

Projective Texture Mapping with Full Panorama

3D graphics, raster and colors CS312 Fall 2010

MODELING AND HIERARCHY

An Animation Synthesis System based on 2D Skeleton Structures of Images

CS4670/5760: Computer Vision Kavita Bala Scott Wehrwein. Lecture 23: Photometric Stereo

Image warping and stitching

Computer Graphics. Bing-Yu Chen National Taiwan University The University of Tokyo

Multi-view Stereo. Ivo Boyadzhiev CS7670: September 13, 2011

Image-Based Rendering and Modeling. IBR Approaches for View Synthesis

A virtual tour of free viewpoint rendering

Neue Verfahren der Bildverarbeitung auch zur Erfassung von Schäden in Abwasserkanälen?

Image or Object? Is this real?

A SXGA 3D Display Processor with Reduced Rendering Data and Enhanced Precision. Seok-Hoon Kim MVLSI Lab., KAIST

Image Warping and Mosacing

Image-Based Modeling and Rendering

Computer Graphics. Bing-Yu Chen National Taiwan University

Synthesizing Realistic Facial Expressions from Photographs

Multiple View Geometry

Scene Management. Video Game Technologies 11498: MSc in Computer Science and Engineering 11156: MSc in Game Design and Development

CS 112 The Rendering Pipeline. Slide 1

Graphics Hardware and OpenGL

Warping, Morphing and Mosaics

INTERNATIONAL ORGANISATION FOR STANDARDISATION ORGANISATION INTERNATIONALE DE NORMALISATION ISO/IEC JTC1/SC29/WG11 CODING OF MOVING PICTURES AND AUDIO

Volumetric Scene Reconstruction from Multiple Views

Motus Unitatis, an Animation Editor

CS 498 VR. Lecture 20-4/11/18. go.illinois.edu/vrlect20

Input Nodes. Surface Input. Surface Input Nodal Motion Nodal Displacement Instance Generator Light Flocking

Lecture 15: Image-Based Rendering and the Light Field. Kayvon Fatahalian CMU : Graphics and Imaging Architectures (Fall 2011)

CS Exam 1 Review Problems Fall 2017

CS 563 Advanced Topics in Computer Graphics Camera Models. by Kevin Kardian

Multiple Views Geometry

DEPTH, STEREO AND FOCUS WITH LIGHT-FIELD CAMERAS

CS770/870 Spring 2017 Animation Basics

CS770/870 Spring 2017 Animation Basics

CS4495/6495 Introduction to Computer Vision. 3B-L3 Stereo correspondence

COMPUTER GRAPHICS COURSE. Rendering Pipelines

Multi-View Stereo for Static and Dynamic Scenes

Splines. Connecting the Dots

Transcription:

Motivation Rendering from Images Image rocessing: View and CS334 Given left image right image Create intermediate images simulates camera movement [Seitz96] Related Work anoramas ([Chen95/QuicktimeVR], etc) user can look in any direction at few given locations but camera translations are not allowed Overview View morphing image pre-warping image morphing image post-warping Quicktime VR Demo Overview Examples View morphing image pre-warping image morphing image post-warping 1

Identify correspondences between input/output image roduce a sequence of images that allow a smooth transition from the input image to the output image 0 k n 2. Linear interpolation k n k n k = (1 ) 0 + n frame 0 frame k frame n 2

is not shape preserving Overview View Morphing Examples View morphing image pre-warping image morphing image post-warping View Morphing Shape preserving morph Three step algorithm rewarp first and last images to parallel views Image morph between prewarped images ostwarp to interpolated view arallel views same image plane image plane parallel to segment connecting the two centers of projection 0 n rewarp np compute parallel views, p rotate and to parallel views prewarp correspondence is ( 0, n ) -> ( op, np ) p 3

Step 3: ostwarping 0 n Shape preserving Use prewarped correspondences Interpolate from ostwarp morphed image create intermediate view is known interpolate view direction and tilt rotate morphed image to intermediate view kp np p View morphing View morphing View morphing is shape preserving View Morphing More Examples Using computer vision/stereo reconstruction techniques Overview View morphing, more details for synthetic rendering image pre-warping image morphing image post-warping 4

arallel views arallel views use for x (a p vector) 0 n 0 n np np p p arallel views use for x (a p vector) use (a 0 x b 0 ) x (a n x b n ) as y (-b p ) arallel views use for x (a p vector) use (a 0 x b 0 ) x (a n x b n ) as y (-b p ) use z = x cross y 0 n 0 n np np p p arallel views use for x (a p vector) use (a 0 x b 0 ) x (a n x b n ) as y (-b p ) use z = x cross y 0 use same pixel size n use wider field of view 0 n prewarping using reprojection of rays np np p p 5

0 n prewarping using reprojection of rays look up all the rays of the prewarped view in the original view prewarping using reprojection of rays look up all the rays of the prewarped view in the original view 0 alternative: prewarping using n texture mapping np np p p prewarping using reprojection of rays look up all the rays of the prewarped view in the original view 0 n np alternative: prewarping using texture mapping create polygon for image plane consider it texture mapped with the image itself render scene from prewarped view I if you go this path you will have to np implement clipping with prewarped plane note: texture mapping in hardware 0 prewarping correspondences for all pairs of correspondence 0 n n np p 0 n prewarping correspondences for all pairs of correspondence 0 n project 0 on, computing project n on p, computing np prewarped correspondence is op np 0 n use prewarped correspondences to compute a correspondence for all pixels in linearly interpolate to intermediate kp np np p p 6

use prewarped correspondences to compute a correspondence for all pixels in linearly interpolate to 0 intermediate n use prewarped correspondences to compute a correspondence for all pixels in linearly interpolate to 0 intermediate n kp np p corresponding pixels are on same line in prewarped views kp np corresponding pixels are on same line in prewarped views preventing holes p use prewarped correspondences to compute a correspondence for all pixels in I I 0p 0 linearly interpolate to intermediate 0 kp n corresponding pixels are on same line in prewarped views preventing holes or linearly interpolate between np I consecutive samples np or postprocess morphed image looking for background pixels and C replacing them with neighboring k values use prewarped correspondences to compute a correspondence for all pixels in linearly interpolate to intermediate corresponding pixels are on same line in prewarped views preventing holes or linearly interpolate between consecutive samples or postprocess morphed image looking p for background pixels and replacing them with neighboring values visibility artifacts collision of samples 0 in prewarped views kp n np use prewarped correspondences to compute a correspondence for all pixels in linearly interpolate to intermediate corresponding pixels are on same line in prewarped views preventing holes or linearly interpolate between consecutive samples or postprocess morphed image looking p for background pixels and replacing them with neighboring values visibility artifacts collision of samples Zbuffer of disparity 0 in prewarped views kp n np use prewarped correspondences to compute a correspondence for all pixels in linearly interpolate to intermediate corresponding pixels are on same line in prewarped views ti h l or linearly interpolate between consecutive samples or postprocess morphed image looking for background pixels and replacing p them with neighboring values visibility artifacts collision of samples Zbuffer of disparity holes 0 preventing holes kp n np 7

Step 3: ostwarping use prewarped correspondences to compute a correspondence for all pixels in linearly interpolate to intermediate corresponding pixels are on same line in prewarped views ti h l or linearly interpolate between consecutive samples or postprocess morphed image looking for background pixels and replacing p them with neighboring values visibility artifacts collision of samples Zbuffer of disparity holes morph p to I kp use additional views 0 preventing holes kp n np create intermediate view is known current view direction is a linear interpolation of the start and end view directions current up vector is a linear interpolation of the start and end up vectors rotate morphed image to intermediate view same as prewarping 8