Audio and video technology

Photo: Prof. Dr. rer. nat. Stefan Bischoff
Ihr Ansprechpartner
Prof. Dr. rer. nat.
Stefan Bischoff
Faculty of Electrical Engineering and Computer Science
02763 Zittau
Schwenninger Weg 1
Building Z VII, Room 413
+49 3583 612-4868
+49 3583 612-4819

Winter semester: Lecture:   2 SWS
Final performance: Oral examination
Admission requirements for the examination:
Successful participation in the course Audio-Video Technology

Pictures, tables, diagrams, etc. in pdf format

Introduction

  • Physiology of hearing and seeing, image and sound signals
  • Data acquisition and representation in the computer
  • Spatial discretization, quantization, colour and time discretization
  • Color spaces (RGB, CMYK, YUV, HSV)
  • Mathematical description

Pre-processing I

  • Gray value scaling
  • Geometric correction
  • Camera calibration

Preprocessing II

  • Convolution in the time and frequency domain
  • Mask and filter techniques, denoising
  • Morphological operations
  • Differential operators, edge extraction (edge detection)

Compression I

  • Lossless and lossy compression
    BMP, JPEG, JPEG2000

Compression II

  • Video compression methods
  • MPEG 4, H264
  • Audio compression
  • MP3 COMPRESSION

Segmentation I

  • Histogram-based methods
  • Point-, edge- and region-oriented methods
  • Knowledge-based methods, Hough transformation
  • Template matching, active contour snakes

Segmentation II

  • Motion segmentation
  • Background subtraction, optical flow
  • Speech segmentation

Feature extraction

  • Color, texture and shape descriptors
  • Principle Component Analysis (PCA)
  • Video descriptors
  • Audio descriptors

Classification I

  • Classification basics
  • Prototypes, cluster analysis
  • Distance-based classification: KNN
  • Statistical methods
  • Bayesian classifier

Classification II

  • Speech recognition
  • Hidden Markov Model
  • Dynamic programming, Viterbi algorithm

Adaptive classification

  • Machine learning methods
  • Supervised and unsupervised learning
  • Artificial neural networks, Perceptron, MLP
  • Support Vector Machine (SVM)

Multi-sensor technology

  • Multi-sensor fusion to increase recognition rates
  • TOF depth sensors
  • Stereo, 3D scene reconstruction
  • 3D glasses, anaglyph, 3D shutter glasses
  • 3D displays

Important applications: Optical quality control in production Robotics, automotive sector Medical diagnostics Video conferencing systems Biometrics Security