CSC4140代做、代写Python/Java编程设计

日期：2024-05-15 08:45

CSC4140 Final Projects

April 27, 2024

The nal is 40% of the total mark.

We encourage to help eachother but do not show the same thing in your report and do not cheat!

Strict Due Date: 11:59PM, May 20th, 2024

Student ID:

Student Name:

This assignment represents my own work in accordance with University regulations.

Signature:

1

1 Transient Rendering through Scattering Medium

Problem Description

Participating media are used to simulate materials ranging from fog, smoke, and clouds, over

translucent materials such as skin or milk, to fuzzy structured substances such as woven or knitted

cloth. Participating media are usually attached to shapes in the scene. When a shape marks the

transition to a participating medium, it is necessary to provide information about the two media

that lie at the interior and exterior of the shape. This informs the renderer about what happens

in the region of space surrounding the surface. In many practical use cases, it is sucient only

to specify an interior medium and to assume the exterior medium (e.g., air), not to inuence the

light transport.

Transient rendering is proposed to simulate how the light propagates in the space. Instead of

the traditional renderer, it assumes the light speed is limited. Transient rendering for participating

media helps to provide a new simulation tool to achieve a new sensing technology in extreme

weather condition.

Goals and Deliverbles

Based on the code of Dierentiable Transient Rendering linked below, realize a renderer in the

FOG medium. You can refer to any renderer or code for participating media. For this project,

deliver a series of transient images generated with the Dierentiable Transient Renderer.

When working on this project, you will have to gure out how to embed the participating media

into the given engine.

Resources

1. Dierentiable Transient Rendering

2. Code

3. Mitsuba

2 Realize BDPT (use cuda)

Problem Description

Based on the code of assignment 5 and 6, realize your own Bi-Directional Path Tracer.

2

Goals and Deliverbles

Render the given scenes in assignments 5 and 6 using your own BDPT and compare the difference with the current one in your report. We encourage you to use CUDA to implement them

to avoid the hours-long rendering process. CUDA has become very easy, and it's just a library for

parallel computing and rendering.

Find a caustic scene like a lens or glass ball, render it with BDPT and the path tracer, compare

the results, and explain.

Resources

M. Clark, "CUDA Pro Tip: Kepler Texture Objects Improve Performance and Flexibility",

NVidia Accelerated Computinig, 2013. [Online]

T. Karras, "Thinking Parallel, Part III: Tree Construction on the GPU", NVidia Accelerated

Computing, 2012. [Online]

T. Karras, "Thinking Parallel, Part II: Tree Traversal on the GPU", NVidia Accelerated Computing, 2012. [Online]

E. Veach, "Robust Monte Carlo Methods for Light Transport Simulation", Ph.D, Stanford

University, 1997.

3 Realize Spectral Ray Tracing and Learn to Use "Nvidia

OptiX"

Problem Description

The current implementation of the raytracer cannot model dispersion and chromatic aberrations

because its light model is not wavelength-dependent. Currently, indices of refraction are constant

rather than dierent for each wavelength. You can implement your code based on assignment 6.

Nvidia OptiX is a high-level GPU-accelerated ray-casting API. If your computer supports

Nvidia RTX, We strongly recommend you try it and coding based on this API instead of the code

of assignment 6.

Goals and Deliverbles

Implement spectral ray tracing by tracing rays of dierent wavelengths sampled using the human

eye's wavelength prole for each color (RGB). By modeling dierent indices of refraction based on

those wavelengths for glass-like materials, we hope you to reproduce eects such as the dispersion

of light through a prism, the changing colors based on the viewing angle for a lens on a reective

3

surface with a thin lm (such as a DVD), as well as model chromatic aberrations present in real

camera systems with lenses. Additionally, it would be best if you created wavelength-dependent

bsdfs and lighting. We hope to model dierent temperature lights.

1. Prism scene rendering

2. Disk/bubbles scene rendering (Add dierent environment maps (potential source from Light

Probe Library). Images from the light probe library are in HDR format, suitable for spectral

ray-tracing since you have the more realistic spectrum distribution of each scene pixel).

3. Correctly simulates chromatic aberration of dierent lenses.

4. Compare rendered images with real photos we take of the objects (e.g., disk).

5. Compare rendering under dierent temperature lights.

6. Finally, we hope you deliver a synthesized image that harmoniously combines objects that

best illustrate the eectiveness of our spectral ray tracer. (e.g., gemstones, etc., suggestions

on this would be helpful!)

7. Optional: Add fog/volumetric scattering so that rainbows can be seen.

Tasks:

1. Change lenstester to also include wavelength argument that the user can set. (mainly for

debugging purposes)

2. Refactor code so that rays have a wavelength argument that can be passed in and checked

as well as that functions that return Spectrums now return a single intensity value

3. Change raytrace_pixel to ask for multiple ray samples for each color channel, then combine

those color channels

4. Change camera.generate ray to take in a color channel argument and sample that color

channel's wavelength distribution (Gaussian) to change the ray's wavelength

5. Change lens_camera's tracing through the lens to use the wavelength argument to change

indices of refraction when tracing through the lens

6. Change sample_L of lights to have a wavelength-dependent intensity to simulate dierent

colors of lights (maybe initialize lights with a temperature argument and model them as ideal

black bodies to get the intensities for each color)

7. Rewrite BSDFs of colored objects to return a wavelength-dependent magnitude instead of a

constant spectrum argument.

4

8. Rewrite/write glass BSDF to have wavelength-dependent indices of refraction (similar code

as lens_camera's tracing)

9. Write a bubble/ thin-lm interference BSDF that uses wavelength, thickness, and light to

determine if the interference occurs (integer multiples of wavelength)

10. Write new scene/dae les (using Blender )/mess with the parser to create a triangular prism

would want a small area of light create a disk + reective surface + transparent coating

Resources

1. AN INTRODUCTION TO NVIDIA OPTIX

2. Prisms and Rainbows: a Dispersion Model for Computer Graphics

3. Iridescent Surface Rendering with Per-channel Rotation of Anisotropic Microfacet Distribution

4. Rendering Iridescent Colors of Optical Disks

5. Derive spectrum from RGB triple

6. soap bubbles 1

7. soap bubbles 2

Other useful links: [1] refractive index [2] refractive indices [3] glassner [4] hyperphysics [5]

Morris, Nigel. "Capturing the Reectance Model of Soap Bubbles." University of Toronto

(2003).

4 Smooth Mesh Estimation from Depth Data using NonSmooth Convex Optimization

Problem Description

Meshes are commonly used as 3D maps since they encode the topology of the scene while being

lightweight. Unfortunately, 3D meshes are mathematically dicult to handle directly because of

their combinatorial and discrete nature. Therefore, most approaches generate 3D meshes of a scene

after fusing depth data using volumetric or other representations. Nevertheless, volumetric fusion

remains computationally expensive both in terms of speed and memory. The main references for

this project are this paper on the Smooth Mesh Estimation from Depth , and you can use the

attached test data.

5

Goals and Deliverbles

Your task is to implement the given paper. To simplify the task, you are allowed to use any

package available to help you dealing with the optimization part. Or you can grab a certain piece

of code block to help you complete this task.

Final Note

You have achieved a milestone in Computer Graphics. Here your task left is to make some

fancy results and reports! Computer Graphics is not only a science of producing graphical images

with the aid of a computer but also a fancy art! Again, always be creative!

6