Towards a wearable system for continuous monitoring of sniffing and panting in dogs

摘要: Although numerous advances have been made in instrumental odor detection systems, these still cannot match the efficient sampling, discrimination, agile mobility and olfactory acuity of dogs. A limiting step using dogs to detect odors is subjectivity translation information processed by dog its handler. We present our preliminary efforts towards a wireless wearable system for continuous auscultation respiratory behavior recording internal sounds at neck chest means commercially available electronic stethoscope provide objective decision support handlers. identified discrete features sniffing panting time domain utilize event duration, rate, mean energy, number consecutive events row build tree classifier. Since feature extraction requires segmentation individual events, we developed an adaptive method short-time energy contour threshold. The performance was evaluated on recordings from Greyhound Labrador Retriever achieved high classification accuracies.