EC4480 Image Processing and Recognition

This course provides image processing background for understanding modern military applications, such as long range target selection, medium range identification, and short range guidance of new weapons systems. Subjects include image sampling and quantization, image representation, enhancement, transformation, encoding, and data compression. Predictive coding, transform coding, and inter frame coding techniques are also introduced. 3D to 2D imaging projections are also introduced to extract 3D information either from motion or stereo imaging. Some effort is directed toward image compression techniques particularly suited for multimedia video conferencing.

Prerequisite

EC3400

Lecture Hours

3

Lab Hours

2

Course Learning Outcomes

·       2D Fourier Transforms, Discrete Fourier Transform (DFT).

·       2D Filter design and the McClellan transformation.

·       Efficient representation of signals and image compression.

·       Discrete Karhunen-Loeve Transform (DKLT) and Discrete Cosine Transform (DCT).

·       Multi Resolution decomposition, in 1D and 2D domains.

·       Quadrature Mirror Filters, Conditions for Perfect Reconstruction and non-Aliasing.

·       Biorthognal and Daubechies Filters, and application to compression and filtering.

·       Video processing:Optical Flow and Kanade Lucas Tomasi (KLT) for motion det.

·       Statistical models for detection for abnormal behavior: Markov chains.

·       Multiscale Representation of Images, Laplacian of Gaussian (LoG) and its properties.

·       SIFT and SURF algorithms for Object Recognition.