The variable cost of producing market hogs is largely centered around feed cost, with diet formulation and feed conversion being the two largest contributors. Feed cost has been estimated to account for 60 to 70% of production costs, with efforts to reduce it being at the forefront of competitive systems [
1]. Diet cost can be manipulated in many ways, such as utilizing alternative feed ingredients, replacing traditional protein sources with crystalline amino acids, altering feed particle size, and/or through enhanced management of animals and their environment. Altogether, the inclusion of dried distiller grains with solubles (DDGS; a major co-product from the production of ethanol from grain) is one of the most common ways to mitigate feed costs [
2]. However, inclusion rates above 15% of the diet can have diminishing returns towards efficiency [
3,
4]. Feed conversion ratio (FCR) is commonly described by the proportion of average daily feed intake (ADFI):average daily gain (ADG) with improvements being made through genetic selection, management practices, and nutrition. An increase in FCR has traditionally been an area of focus for genetic improvement. Unfortunately, diminishing returns have been observed in response to the intense selection of FCR due to underlying correlations between growth and carcass traits [
5]. For example, Godinho et al. [
6] reported negative genetic correlations between FCR and ADG (−0.32) and between FCR and protein deposition (−0.47). Additionally, FCR showed positive genetic correlations between backfat depth (BF) (+0.16) and average daily feed intake (ADFI) (+0.54) [
6]. Furthermore, roughly a third of the phenotypic variation associated with ADFI is accredited to residual feed intake (RFI) [
7,
8]), making the accurate prediction of growth performance and potential genetic selection difficult. Residual feed intake is defined as the feed consumed above expected requirements for production and maintenance [
9]. Genetic selection for lower RFI has been emphasized to create offspring that consume less feed without sacrificing growth [
10]. Selection for increased feed efficiency based on RFI results has been shown to improve efficiency in the grow-finish stages of production but is suggested to be combined with the growth rate in the selection scheme in order to be applicable to producers [
11]. Without this integral step, producers would most likely observe decreases in growth performance due to unfavorable associations. Factors affecting RFI include digestion efficiency, metabolism, maintenance, and activity. Each of these factors has been shown to have a high level of variability among individuals [
12,
13]. Specifically, maintenance represents a significant portion of the daily energy intake in a pig, with roughly 34% of the daily energy intake of a 70 kg pig being directed to maintenance [
14]. Typically, the activity component is disregarded due to the lack of technology to objectively measure it. To be more precise in estimating expected feed intake and to more accurately select for RFI associated traits, further understanding is needed in the area of calorie expenditure due to activity patterns.
In the past, swine production systems relied solely on human observation for monitoring behavior and activity trends within group-housed animals, with activity being a secondary concern in most circumstances [
15]. Most of these observational times are dedicated to health and welfare, with caretakers focusing on behaviors that can be indicative of injury, sickness, and/or stress. It has been estimated that in modern commercial farms, human observation per animal each day can be as low as only a few seconds [
16]. Even if more human interaction/labor were dedicated to observing general behavior patterns that may indicate activity levels, the presence of a human can significantly impact the adaptive behavior of the pigs [
17]. Furthermore, human observation is often heavily subjective, with technicians varying in their observations. Detection of changes in behavior and activity patterns of individual pigs housed in groups is crucial to maintaining animal health and wellbeing standards. Such changes can be indicative of internal/external stressors impacting the pig. Animal-based indicators offer the most centralized measure of animal wellbeing [
18]. Rapid detection of such indicators may lead to faster response time in treatment of unwell animals.
One developing solution to overcome the constraints of human observation is the progression of precision livestock technology (PLT). A primary objective with PLT is to develop a real-time/on-line system that has capabilities to maintain the individual animal ID, track traits of interest on an individual basis, and provide accurate predictions of variable change [
19]. Not only will PLT allow producers to more accurately measure and assess key traits of interest, PLT can be used to improve the health and well-being of the herd. Advancements have shown that such technologies may be able to identify an animal in need of treatment prior to human observation. In the beginning, attempts to solve these PLT goals included attaching a wide array of electronic devices to individual animals, such as radio-frequency identification (RFID) collars and tags [
20,
21,
22]. Unfortunately, these devices not only require significant maintenance and financial commitment to utilize, but also can be invasive to the animal, creating a welfare concern. Thus, recent approaches are gravitating toward non-invasive, vision-centered methods [
23]. Recent developments in advanced computer vision systems, such as NU
track (NU
track Livestock Monitoring), now allow for activity traits to be quantified and tracked individually in a traditional commercial setting [
24]. NU
track is a fully convolution machine learning program for the long-term location and activity monitoring of individual pigs [
17]. Therefore, the aim of this study was to utilize NU
track to objectively measure nursery and finisher activity levels of pigs from sires with different estimated breeding value (EBV) combinations for growth and feed intake. The EBV is the calculation of the animal’s genetic worth based on performance records of itself and other relatives and is reported as a deviation from the population mean [
25]. Additional objectives included determining variation across activity traits of all animals as well as evaluating the impact of sex on performance and activity traits.