Fun Projects

BRAINS Constellation Detector is an automated detector designed for estimating landmark constellation in 3D digital MR brain images. The tool is also one of a series in BRAINS tool suite for a comprehensive brain image analysis. It has been developed as part of my MS thesis (download here) at the Psychiatric Iowa Neuroimaging Consortium in University of Iowa Hospitals and Clinics, under the supervision of Dr. Hans J Johnson.

Method Highlights:

• Spatial localization using evolutionary principal components
• Image representation and local search using statistical shape models
• Atlas construction using generalized splines
• Eye detection using probabilistic radial Hough transform
• Finding optimal rigid transform using Powell's optimizer
  
• Image clustering using Otsu thresholding
• Landmark search center prediction using morphometric constraining

Demo:
Detection result


GUI tool


Atlas construction result




Reference:
Project homepage at NITRC
My MS Thesis pdf
Defense Presentation pdf

Network based Smartphone client application for viewing online photos

Features:

• Three-axis gyro sensor, and multi-touch technique enabled
• Showing photos on geo-map
• User can scroll, zoom in/out a photo
• Viewing history is presented

Demo:


Reference: iPhone Application Development course presented by Stanford University

Idea: This Flash game encourages players to collaborate with each other to battle with the space invaders with their armed spacecrafts.

Demo:

Idea:

• Feature extraction and matching: SIFT
• Matching points refining: RANSAC
• Model estimation: Linear homography
• Blending: Multiband blending

Demo:



Documentation:
pdf

Idea:

• 2D: Sequential Thinning Method
• 3D: Parallel Thinning Method
• With node (in red) and leaf (in yellow) detection

Demo:



Documentation:
pdf

Requirements:

• Handle reflectivity (via recursive ray firing)
• Handle transparency, with refraction (via recursive ray firing)

Demo:

Idea: The recognition method is based on Local Binary Patterns (LBP) of Quaternionic Gabor features (QGF). By introducing Quaternion Gabor analysis into image representation, we make full use of the interrelationship among different color channels to enhance the performance of the face recognition system. Moreover, the QGF are used to encode the positions and attributes of the face elements. Non-parametric transformation is then imposed on these QGF using LBP method to obtain the robustness against variations of pose, illumination and facial expressions.

Demo: Framework of the method


Publication:
"Local Quaternionic Gabor Binary Patterns for color face recognition" by Wei Lu; Yi Xu; Xiaokang Yang; Li Song
pdf

Idea: Each block of two co-registered images are decomposed into high frequency and low frequency components. Each block of the fused image is reconstructed from the average of the two low frequency components and the high frequency component with larger energy in the corresponding blocks of input images.

Demo: (Input images in the first two rows have different focuses; input images in the third row have different spectral and spatial resolutions)


Transformation:
Quaternion wavelets
Ref: E. B. Corrochano, "The theory and use of the quaternion wavelet transform," Journal of
Mathematical Imaging and Vision, Vol.24, pp.19-35, 2006.
Contourlets
Ref: M.N. Do and M. Vetterli, "Contourlets," Beyond wavelets, pp.1-27, 2001

Publication: "QWT: Retrospective and New Applications" by Yi Xu, Xiaokang Yang, Li Song, Leonardo Traversoni and Wei Lu
Link