Year: 2017
Type: conference-paper
Abstract: Separation anxiety in dogs is a clinical condition in which a dog develops an intense dependence on their caretaker. It manifests itself in maladaptive physical and psychological behaviors that are similar to that of panic attacks in humans when the dog is isolated from its owner. These include excessive vocalizations, destructive behaviors, and in more extreme cases, self-mutilation. Currently, there are no direct ways to diagnose separation anxiety without the aid an owner’s report. To ease this diagnostic process, it would be useful to utilize a tool that can distinguish between certain vocalization patterns in dogs. Literature suggests that some vocalizations, such as whining, reveals information about the dogs wellbeing. The purpose of this study is to initiate steps to build and implement a classifier that will distinguish between different dog vocalizations. This process involved meticulously preparing the data by extracting, labeling, and separating the vocalizations in a sound editing software. Mathematical models of several acoustic parameters were constructed to obtain critical information from each file. With this information, a classifier can be built and could ultimately aid in understanding the underlying emotions in dogs suffering from separation anxiety.
Authors: Karina Sequera, Nan Kong, Niwako Ogata
Language: English
Subfield: Genetics
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W2749662876