Psych221 Project Suggestions

These are project suggestions for Psych 221. We update this page regularly with ideas for projects. To see project examples from previous years, visit SCIEN Class Projects Page.

Projects 2012

Visibility of movie subtitles

A persistent problem in watching foreign movies is that sometimes the subtitles are illegible. Why? Because the contrast of the default background that is assumed is wrong and you have white characters on a light background. I assume this is done automatically because it is too expensive to have people judge frame by frame whether the script is visible. Need I say more. Some automated system that could assess the brightness of the standard background space where subtitles are printed and then adjust the contrast to be legible would be a huge improvement for the industry.

E. Markman, a committed viewer of subtitled films.

3D Distributed Ray Tracing calculations

PBRT

Radiance

RenderToolbox

Photo retouching metric (Kee and Farid)

A perceptual metric for photo retouching Eric Kee and Hany Farid

Department of Computer Science, Dartmouth College, Hanover, NH 03755 October 19, 2011 (received for review July 5, 2011)

In recent years, advertisers and magazine editors have been widely criticized for taking digital photo retouching to an extreme. Impossibly thin, tall, and wrinkle- and blemish-free models are routinely splashed onto billboards, advertisements, and magazine covers. The ubiquity of these unrealistic and highly idealized images has been linked to eating disorders and body image dissatisfaction in men, women, and children. In response, several countries have considered legislating the labeling of retouched photos. We describe a quantitative and perceptually meaningful metric of photo retouching. Photographs are rated on the degree to which they have been digitally altered by explicitly modeling and estimating geometric and photometric changes. This metric correlates well with perceptual judgments of photo retouching and can be used to objectively judge by how much a retouched photo has strayed from reality.

• Attend the talk by Farid at Stanford, Jan. 31.
• Implement and analyze the algorithm.
• Perform experiments with existing online pictures.

Suggest improvements

3D Image Quality Metrics

Develop algorithms for Shooting in 3D and Displaying in 2D. Explore ways in which to improve 2D rendering of 3D content in order to enhance “immersive video”.

MRI possibilities (Aviv)

RF excite inhomogeneity estimations. RF receive inhomogeneity estimations. T1 and MO fits engineering project:

RF excite inhomogeneity measurement 1

The excite inhomogeneity is mesure by EPI T1 mapping. The EPI is register linearly with semi manually processe to the SPGR images. Task: define a nonlinear solution that bring the EPI map closer to the SPGR maps. This non linear registration may be fine as it is registration between to images of the same brain (and not as typically done to an MNI space). The idea is to try the available method and find the best one.

RF excite inhomogeneity measurement 2

The excite inhomogeneity is mesure by EPI T1 mapping. The EPI excite inhomogeneity map only cuver part of the SPGR image, further more the excite inhomogeneity is noise in nature (short TR's EPI's), and not perfectly register (Project 1). The noisy area as well as the gap in the map are filled with the idea the the excite inhomogeneity is smooth. to do so we use both local and global regressions. task: define the most robust and accurate something function to estimate the excite inhomogeneity.

RF receive inhomogeneity measurement 3

RF receive inhomogeneity measurement: - (any result in here can be also relevant for parallel imaging- i do have multi coils assest data for it) Project 1: The receive inhomogeneity is calculate using the multi coil information. The idea is to find the brain image that is common to all the different coils and subtract the coil inhomogeneity. Task: test if the current minimization function is optimal can is there more robust ways to do it? are there still coil inhomogeneity residual that left in the PD map (common bias to coil) is there away to control it?

Project 2

The Brain PD is define as the part that is common to the different coils. by this way we essentially average the PD part of the different coils. The averaging may reduce the signal to noise and it possible that a better avreging is possible with some weights on the different coil signal. there is a  research question in htere becouse the weights might also increase the effect of residual bias of the different coils. 

Task: find the best way to combine the multi coil PD information

T1 and M0 fit

Project 3: we fit the T1 and M0 by a LSQ process in a voxel by voxel way. we use 4 flip angle SPGR and fit the SPGR signal equation with a known excite inhomogeneity map. for intuition T1 map explain the difference between the different SPGR and the M0 is the common part of the different SPGR images. other method were suggested to fit the equation like a linearisation of the equation. This method is faster and in my hand also more robust. In the literature this method shown to biased the fitted T1 to high T1 values and suggest to be biased differently for different T1 values. In the literature there was a suggestion to overcome this problem (Nikola and Deioni using this alternative method) by adding weights on the different flip angles interactively. my feeling is that the current LSQ method we use fit the data in a way that increase the noise in the M0 (PD) maps but not in the T1 (it is like the noise is pushed to the M0 part). Task: find the optimal T1 M0 fitting giving the noise in the data, so the fits will be robust and repeatable.

Projects 2011

Software

Camera Design

Imagine that you work for a cell phone company and you have been assigned the task of selecting the components for a camera that will be built into the phone. Because you work for a large vendor, you have parts suppliers calling you every day encouraging you to purchase their devices. Typically, these suppliers offer one component and it is your job to assemble a large variety of components into a working system. In this project you will be asked to develop a method to make these purchasing decisions. Specifically, you will have to decide on which lens, anti-aliasing and IR filter, sensor, and color filter array you would like to use for your system.

For more details see: Camera Design

Processing algorithm analysis

We provide noisy sensor images and desired sRGB renderings of those images. Your implement processing algorithms (color transform(s), denoising, demosaicking, and display rendering steps) to approximate the high quality desired sRGB renderings. This can (and should) be a team project with students who implement different parts of the image processing pipeline. We evaluate your methods by providing new test images for your pipeline and evaluate the quality of the sRGB rendering.

For more details see: Pipeline Analysis.

Hardware

Build an LED multispectral light source

We built one, so we know it can be done. We have a part list and plans. We have ideas for the next generation. Do you like to build stuff? This is your chance to do a hardware project.

Psych 221 Projects 2010

Web page of Project Ideas for 2010

Psych 221 Projects 2008

PDF of Project Ideas in 2008