Foundation in Image and Signal Processing

The aim of the module is to teach basic techniques of classical image and signal processing. The methods covered build on procedures from stochastics and analysis.

Table of contents

• Signal and image acquisition and their storage
• Pattern recognition based on correlation in signals and images
• Convolution and filtering in images and signals (linear/non-linear filtering, edge detection)
• Keypoint approaches:feature detectors and descriptors for images

Students will be familiar with the processes and formats of digital signal and image generation, and will be able to use classical techniques. This includes:

• Images in Python, libraries PIL/Pillow, OpenCV and skimage
• Image representation: color spaces, color planes, image formats
• Histograms for signals and images: Creating, analysis, limit methods for binary segmentation, white balance.
• Morphological operations for signals and images: binary images, grayscale images

Students understand the principle of convolution and its use in linear and non-linear filtering for signals and images. Methods of edge and corner detection such as Canny edge detector, Harris corner detector, Hough transform as well as filtering in the frequency domain e.g. by means of fast Fourier transform can be applied to signals and images.

Students will be able to detect keypoints of feaetures in images using feature descriptors such as FAST, HoG, SIFT, or BRISK.