A SOUND ANALYSIS SYSTEM BASED ON LABVIEW® APPLIED TO THE ANALYSIS OF SUCKLING GRUNTS OF DOMESTIC PIGSSUS SCROFA

Sows' Responses to Piglets in Distress: An Experimental Investigation in a Natural Setting

Domestic pigs (Sus scrofa) possess complex socio-cognitive skills, and sows show high inter-individual variability in maternal behaviour. To evaluate how females-reared under natural conditions-react to the isolation calls of their own piglets or those of other females, we conducted observations and experimental trials. In January-February 2021, we conducted all-occurrences sampling on affiliation, aggression, and lactation (daily, 7:30-16:30 h) on six lactating and four non-lactating females at the ethical farm Parva Domus (Turin, Italy). The trials (30 s each, n = 37/sow) consisted of briefly catching and restraining a piglet. We recorded the sow response (none/reactive/proactive movement towards the piglet; self-directed anxiety behaviours such as body shaking) before and during the trial and under control conditions. Increased levels of anxiety behaviour in sows were accompanied by an increased frequency of responses. Less aggressive sows and lactating sows showed the highest frequencies of response. Finally, the isolation calls' maximum intensity had an influence on the type of response observed, with higher proactive response frequencies following lower intensity isolation calls. Our results suggest that being under lactation could play a key role in increasing sow response levels and that specific acoustic features may influence the response.

On-farm detection of claw lesions in dairy cows based on acoustic analyses and machine learning

Claw lesions are a serious problem on dairy farms, affecting both the health and welfare of the cow. Automated detection of lameness with a practical, on-farm application would support the early detection and treatment of lame cows, potentially reducing the number and severity of claw lesions. Therefore, in this study, a method was proposed for the detection of claw lesions based on the acoustic analysis of a cow's gait. A panel was constructed to measure the impact sound of animals walking over it. The recorded impact sound was edited, and 640 sound files from 64 cows were analyzed. The classification of animal-lameness status was performed using a machine-learning process with a random forest algorithm. The gold standard was a 2-point scale of hoof-trimming results (healthy vs. affected), and 38 properties of the recorded sound files were used as influencing factors. A prediction model for classifying the cow lameness was built using a random forest algorithm. This was validated by comparing the reference output from hoof-trimming with the model output concerning the impact sound. Altering the likelihood settings and changing the cutoff value to predict lame animals improved the prediction model. At a cutoff at 0.4, a decreased false-negative rate was generated, and the false-positive rate only increased slightly. This model obtained a sensitivity of 0.81 and a specificity of 0.97. With this procedure, Cohen's Kappa value of 0.80 showed good agreement between model classification and diagnoses from hoof-trimming. In summary, the prediction model enabled the detection of cows with claw lesions. This study shows that lameness can be detected by machine learning from the impact sound of hoofs in dairy cows.

Discerning pig screams in production environments

Pig vocalisations convey information about their current state of health and welfare. Continuously monitoring these vocalisations can provide useful information for the farmer. For instance, pig screams can indicate stressful situations. When monitoring screams, other sounds can interfere with scream detection. Therefore, identifying screams from other sounds is essential. The objective of this study was to understand which sound features define a scream. Therefore, a method to detect screams based on sound features with physical meaning and explicit rules was developed. To achieve this, 7 hours of labelled data from 24 pigs was used. The developed detection method attained 72% sensitivity, 91% specificity and 83% precision. As a result, the detection method showed that screams contain the following features discerning them from other sounds: a formant structure, adequate power, high frequency content, sufficient variability and duration.

Automated recording of stress vocalisations as a tool to document impaired welfare in pigs

The vocalisations of animals are results of particular emotional states. For example, the stress screams of pigs may be indicators of disturbed welfare. Our objective was to develop a system to monitor and record levels of stress calls in pigs, which could be employed in environments of breeding, transportation and slaughter. Using a combination of sound analysis by linear prediction coding and artificial neural networks, it was possible to detect the stress vocalisations of pigs in noisy pig units with few recognition errors (<5%). The system (STREMODO: stress monitor and documentation unit) running on PCs is insensitive to environmental noise, human speech and pig vocalisations other than screams. As a stand-alone device it can be routinely used for the objective, non-invasive measurement of acute stress in various farming environments. The system delivers reliable, reproducible registrations of stress vocalisations. Its detection quality in commercial systems was found to correlate well with that of human experts. STREMODO is particularly well-suited for comparisons of housing and management regimes. Since the system can be trained to recognise various animal vocalisations, its use with other species is also well within its scope.

Linear prediction coding analysis and self-organizing feature map as tools to classify stress calls of domestic pigs (Sus scrofa)

It is assumed that calls may give information about the inner (emotional) state of an animal. Hence, in the last years sound analysis has become an increasingly important tool for the interpretation of the behavior, the health condition, and the well-being of animals. A procedure was developed that allows the characterization, classification, and visualization of the cluster structures of stress calls of domestic pigs (Sus scrofa). Based on the acoustic model of the sound production the extraction of features from calls was performed with linear prediction coding (LPC). A vector-based self-organizing neuronal network was trained with the determined LPC coefficients, resulting in a feature map. The cluster structure of the calls was then visualized with a unified matrix and the neurons were labeled for their input origin. The basic applicability of the procedure was tested by using two examples which were of special interest for a possible evaluation of the normal farming practice. The procedure worked well both in discriminating individual piglets by their scream characteristics and in classifying pig stress calls vs other calls and noise occurring under normal farming conditions.