This dataset contains the predicted posture of 22 pigs under two contrasted conditions: Heat Stress (HS) corresponding to the tropical climate (between 20.3°C and 27.9°C) and Thermoneutral (TN) consisting of an indoor temperature-controlled room at 22°C. At the beginning of the monitoring, animals were 17 weeks of age and were placed in individual metal-slatted pens (1.60 × 0.84 m) equipped with a stainless steel feeder and a nipple water drinker. The energy and protein ration were distributed throughout the day in three meals : between 6:00 and 8:00 (half of the soybean meal ration), between 9:00 and 11:00 (the totality of the green bananas) and between 12:00 and 14:00 (the other half of the soybean meal). Water was available to all pigs ad libitum from a nipple drinker designed to limit waste. The daily feed intake (ADFI, in g/day) was calculated as the difference between the amount of feed provided on day n and the amount of feed left over the morning of day n+1. Average Daily Gain (ADG, in g/day) was calculated as the difference between pig BW at the end and at the beginning of the trial divided by the number of days of the trial. Feed Efficiency (FE) was calculated as ADG divided by ADFI. We used 8 Time-Lapse cameras (TLC2000 Pro, manufactured in 2018, by the brand Brinno) to monitor the animal at 20-second intervals, from 8:00 to 18:00. Each camera capturing a single animal at a time, cameras were switched every 2 days to allow monitoring of the 11 animals of each trial. On average each animal was monitored for 4.3 days (SD = 1.33 days, min=2, max=8).

More details are available in the associated publication.

The dataset is a table containing 157,279 rows and 8 columns. One row corresponds to one observation, i.e. the posture of the animal at a given time, with the associated meta-data. The first column, date, provides the date and time of the day at which the observation was made. The column anim provides the animal number. Note that this is a fake id and does not correspond to the national identification number. But this can be used to differentiate the observation from the different animals. The column posture provides the estimated posture of the animal. Five postures were considered. Standing: Body in upright position, with extended legs, and when only hooves are in contact with the floor. Sitting: Animal is partly rested on stretched front legs with caudal end of body contacting the floor. Sternal: The animal is lying on abdomen/sternum with the front legs folded under the body. The hind legs folded invisible or stretched out visible. Abdomen is totally/partially obscured. Lateral: The animal is lying on either side with all four legs visible and abdomen is totally visible. The posture was estimated using a neural network. The column muscle_temp provides the muscle temperature, while the column ambient-temp provides the ambiant temperature. The column condition provides the condition at which the animal was housed, either HS or TN. The column adg provides the ADG and the column feed_efficiency provides de FE.