This software is the result of the diploma thesis "Detection of coughing, clapping and tapping" by Cornelius Reuss. The aim of the work was the real-time recognition of the three examined noise classes in audio signals.
The program uses the equipped standard microphone for recording. An adaptive threshold energy detector detects possible acoustic events in real time. Then follows a classification of events in coughing, clapping or knocking. The result is then displayed in a graphical interface with the microphone signal.
The provided program is compiled for Windows and 64bit Linux and can be downloaded here to test for yourself. Qt and Qwt were used for graphical representation. For Windows, the required libraries are included and the program should be able to be executed after unpacking without further changes. In order to run the program under Linux, the required libraries have to be installed by yourself.