Roger Ng  B.Eng.(Hons)(PolyU), M.Sc.(CUHK), M.Phi.(DPU), MBA(CUNY), Ph.D. Candidate (UCLA), MIEEE

Roger obtained his B.Eng.(Hons) degree in Electronic Engineering from The Hong Kong Polytechnic University, M.Sc. degree in Electronic Engineering from The Chinese University of Hong Kong, M.Phi. degree in Information Technology from the DePaul University, Chicago USA and MBA degree from the Columbia University, New York USA. He is now having his Ph.D. study in Information Engineering in the University of California, Los Angeles USA. His current research interests are computer vision, image coding and image compression. He is a member of Institute of Electrical and Electronic Engineers.

Computer Vision Technology

    - 2D and 3D Imaging
    - Discrete transformation of images
    - Edge detection
    - Image segmentation
    - Texture analysis
    - Pattern Recognition
    - Hough transform
    - Neural networks
    - Object recognition

Image Processing and Video Technology

Human visual system

HVSs consists of eyes and brains. Eyes transform light into neural signals and brains process neural signals and extract needed information. We will learn how a HVS is affected by factors such as Adaptation, Mach Band Effect, Spatial Frequency Response and Temporal Properties of Vision.
The physical perception of color is based upon 3 types of cone cells in the retina. We will learn how standard sensitivity curves of these cones were adopted by the CIE for the standard observer and how color is represented.
Camera & Monitor Characteristics; Gamma Correction.

Filtering and transforms

2D linear system theory: 2D Convolution, PSF and its frequency response; 2D Filter design and 2D DFT/FFT;
Kronecker product; 2D transformation, separable 2D transformation; SVD, DCT, KLT;

Image analysis

Detection of edges: Sobel operator, Laplacian operator, Laplacian of Gaussian operator, Canny edge detector.
Detection of contours.

Image enhancement

Design filters which can enhance an image.
One can design and use this simple filter

to enhance this image

to become

Other technique such as histogram equalization.

Image restoration

When an image is captured, it may be distorted due to camera motion, out-of-focus, atmospheric turbulence and noises from sensors or other components. In this course, we will learn methods to remove these distortions, and pros and cons of these methods.
Restoration methods: Unconstrained least square restoration, Constrained least square restoration, Wiener filtering, Regularization method.

Image coding

Basic coding techniques such as optimal quantization, predictive coding, transform coding and subband coding.
Image Coding standards: JPEG and JPEG2000.

Video Coding Standards

H.261, H.263, MPEG1, MPEG2 and MPEG4.