Course Details

Subject {L-T-P / C} : EC6607 : Transform Domain Signal Processing and Applications { 3-0-0 / 3}

Subject Nature : Theory

Coordinator : Prof. Manish Okade

Syllabus

Introduction to Transform Domain: Motivation, Need for domain change, Continuous Time Fourier Transform, Discrete Time Fourier Transform and few applications to filtering, Orthogonal Transforms, Discrete Fourier Transform and few applications in image processing, Discrete Cosine and Sine Transform with applications to JPEG compression, Walsh-Hadamard Transform and its significance with applications, Short-Time Fourier Transform and Time frequency localization concept with applications to audio signals, Continuous Wavelet Transform from all three points of view viz. scaling and translation, sub-band filtering decomposition, lifting based, Discrete Wavelet Transform and its applications to 1-D signals and JPEG-2000 compression application, Haar Transform and applications. Non-kernel based transforms: Karhunen Loeve Transform with applications to image compression, Singular Value Decomposition with applications to image enhancement , Principal Component Analysis with applications to face recognition. Time to time domain transformation: Hilbert transform, Logarithmic Transform. Transforms based on projections and voting: Radon transform and Hough transform.

Course Objectives

Course Outcomes

1. Recognize the advantages of analysing and processing data in the transform domain. <br />2. Compare and contrast between the available transforms along with suitability of a transform for a particular application. <br />3. Analyse the advantages and shortcomings between kernel and non-kernel based transforms.

Essential Reading

Supplementary Reading