Epsilon Archive for Student Projects

Abstract

Early detection of diseases and injuries in animals is crucial for their health and well-being. Early diagnosis can be assisted by objective registration of different types of physical activities or behaviour patterns. Monitoring specific parameters, such as changes in activity levels or habits, could serve as an indicator of underlying health issues. It can be challenging for pet owners to notice subtle changes in those characteristics at an early stage. It becomes more difficult in the case of parameters of a low frequency of occurrence, such as drinking and littering behaviours. Hydration status is extremely important in cats and changes in drinking and littering patterns could be early symptoms of potential disorders, in particular diabetes mellitus.
There is a noticeable increase in owners’ awareness about the physical and mental health of their pets. With a growing demand for higher standards of tools to assess animals’ everyday habits, more technologies are being developed. Activity monitors utilizing accelerometers provide broad and continuous measures of physical activity, that enable remote and non-invasive monitoring of an individual’s actions.
The aim of this study was to validate the registrations of an activity monitor. Specifically, the study aimed to assess the effectiveness of the activity monitor in detecting drinking and littering activities, which might suggest underlying health issues. To monitor these activities, this study used an activity monitor equipped with an accelerometer and attached to a collar. The validity and effectiveness of the activity monitor were established by comparing the measurements obtained from the activity collar to video recordings from the motion sensor camera.
For forty-eight days, activity data on drinking and littering actions were collected from a single adult cat. Descriptive statistics were performed to summarize the main findings of the dataset to obtain key results. From the total of 5989 recordings registered by the motion sensor camera, 671 recordings containing actions of drinking and littering were selected for further analysis. Accordingly, 53 recordings were extracted from the activity monitor. This study found no correlation between the data obtained from the activity monitor and the video observations from the motion sensor camera. Further research is needed to investigate the reasons behind this lack of agreement and to improve methodologies for monitoring feline activities using activity monitors.
Despite underwhelming findings, it should not rule out all potential applications in monitoring feline behaviors, managing health disorders, and promoting overall health remain promising.