Jason Yu-Tseh Chi

Jason Yu-Tseh Chi

CSE Building E445
Department of Computer & Information Science & Engineering
University of Florida

email: ychi@cise.ufl.edu

A Fast Affine Registration Algorithm for Massive 3D Point Sets.

Implementation of a CVPR 2010 paper - "Common Visual Pattern Discovery via Spatially Coherent Correspondences" by Liu and Yan.

• mex CalcAffinityMatrix3.cpp first. This is the code that calculate affinity matrix. It returns matrix A in sparse format and correpondences indices.
• A simple example is in the file "runthisone.m".
• We use vlfeat package to calculate sift.
• There are couple thresholds in the begining of the file algorithm3.m you can/should play with.
• We borrow the file "plotmatches.m" from the old UCLA sift package to plot the results.

Self organizing map.

Matlab source code.
README.

• Run som.m in matlab after downloading the file. Program draw the results every 200 iterations.
• Please see the README file first.

B-Spline snake. Final project of Medical Imaging Analysis.

• Must install opencv.

Curve Sketcher. (Computer graphics - project 1)

• Execute ./run to compile and run the program.
• Drag any control point to change the shape of the curves.
• Dragging a control point with SHIFT key held will change the z coordinate of the control point.
• Press keys 1, 2, 3, or 4 to switch between different rendering mode.

Tetrahedrons. (Computer graphics - project 2)

• Execute ./run to compile and run the program.
• Select a tetrahedron, hold the mouse left button and move the mouse around. Different tetrahedron has different movement.
• User can change camera viewing angle by mouse.
• Press key g and then 'space' key to shoot bullets from an edge of a tetrahedron. Press key o to switch to tonado mode.

Head and face. (Computer graphics - project 3)

Techniques involved:

• OpenGL related: lighting, Nurbs surface, trimming of nurbs surface, texture mapping, and animation.
• Triangular mesh file(obj).
• Quoternion rotation.

Source code.
README.

• Execute ./run to compile and run the program.
• Please see the README file first.

Understanding Support Vector Machine.

• [error als b0s]= SVM_Simple(data,lbl1,testData,lbl2);
  SVM training. Return largrange multiplier and biases.
• [index] = SVMTest(Alphas,b0s,data,testData);
  SVM testing. Given a collection of largrange multipliers and biases, return index of the testData.
• [error Alphas b0s]= SVM_PCA(trainingData,labels1,testData,labels2,eval,evec,means);
  SVM training scheme by using PCA feature.
• [error Alphas b0s]= SVM_Simple(trainingData,labels1,testData,labels2);
  SVM training scheme by using original features and edge histogram feature.

A MIPS simulator. (Computer Architecture - final project)

• Download, decompress and make.
• Run "MIPSsim fibonacii.bin output.txt". A cycle by cycle CPU status will be recorded in the output.txt file.