Sound detection for study room monitoring and evaluation

ABSTRACT In the field of sound recognition, research and study event detection are still active, with vast majority papers focusing mostly on domain speech music. This paper presents discusses a framework for room system. Feature extraction techniques utilized discussed to obtain parametric type representation analysis intelligent homes machine listening systems specifically room. The conduct within category sounds least accurate was door knock, but accuracy 95.00%, currently in is acknowledged as good, making parameters fit detecting surrounding sounds. performance CNN environmental analyzed using that were defined, an overall 96.8%. result promising learning detects can be applied technology innovative environment.