http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Deep Learning-based Bird Sound Recognition System with Data Pre-processing
Taepyo Jung,Hyungjin Jeon,Cheolmin Jeon,Aarion Cook,Alayna Weiss,Minsun Lee,Anthony H. Smith 대한전자공학회 2019 대한전자공학회 학술대회 Vol.2019 No.11
Over the past decade, scarecrows have begun moving from stand-alone scarecrows to scarecrows using Internet of Things (IOT) technology. This study focuses on building a “smart” scarecrow that recognizes and detects birds by bird sounds. We propose a bird sound recognition model based on data pre-processing and Convolutional Neural Network (CNN). The hypothesis has been established that noise elimination through data pre-processing is more effective than not using data pre-processing to recognize bird sounds. The results showed that the overall performance of the bird and non-bird sound classification through the pre-processing system was 79.8%, which was no different from that of a system without data preprocessing. However, the proposed model has a 4.81% improvement in non-bird sound classification performance compared to models without data preprocessing.