Temporal refuges differ between human and natural top-down pressures in a subordinate carnivore

Animals exhibit variation in their space and time use across an urban-rural gradient. As the top-down influences of apex predators wane due to human-driven declines, landscape level anthropogenic pressures are rising. Human impacts can be analogous to apex predators in that humans can drive increased mortality in both prey species and carnivores, and impact communities through indirect fear effects and food subsidies. Here, we evaluate the time use of a common mesocarnivore across an urban rural gradient, and test whether it is influenced by the intensity of use of a larger carnivore. Using multiple camera-trap surveys, we compared the temporal response of a small carnivore, the raccoon (Procyon lotor), to the larger coyote (Canis latrans) at four sites across Michigan that represented a gradient of pressure from humans. We found that raccoon time use varied by site and was most unique at the rural extreme. Raccoons consistently did not shift their activity pattern in response to coyotes at the site with the highest anthropogenic pressures despite considerable interannual variation, and instead showed the stronger responses to coyotes at more rural sites. Temporal shifts were characterized by raccoons being more diurnal in areas of high coyote activity. We conclude that raccoons do partition time to avoid coyotes. Our results highlight that the variation in raccoon time use across the entirety of the urban-rural gradient needed to be considered, as anthropogenic pressures may dominate and obscure the dynamics of this interaction. In an increasingly anthropocentric world, to understand species interactions, it is imperative that we consider the entire spectrum of human pressures that it may occur within.

Recovering Geometric Information with Learned Texture Perturbations

Regularization is used to avoid overfitting when training a neural network; unfortunately, this reduces the attainable level of detail hindering the ability to capture high-frequency information present in the training data. Even though various approaches may be used to re-introduce high-frequency detail, it typically does not match the training data and is often not time coherent. In the case of network inferred cloth, these sentiments manifest themselves via either a lack of detailed wrinkles or unnaturally appearing and/or time incoherent surrogate wrinkles. Thus, we propose a general strategy whereby high-frequency information is procedurally embedded into low-frequency data so that when the latter is smeared out by the network the former still retains its high-frequency detail. We illustrate this approach by learning texture coordinates which when smeared do not in turn smear out the high-frequency detail in the texture itself but merely smoothly distort it. Notably, we prescribe perturbed texture coordinates that are subsequently used to correct the over-smoothed appearance of inferred cloth, and correcting the appearance from multiple camera views naturally recovers lost geometric information.

Effects of Age on Match-related Acceleration and Deceleration Efforts in Elite Soccer Players

The aim of this study was to examine the effects of chronological age on acceleration and deceleration match performance in professional soccer players. A total of 5317 individual match observations were collected on 420 professional players competing in the Spanish LaLiga during the 2018–2019 season, using a multiple-camera computerised tracking system (TRACAB; ChyronHego, Melville, NY, USA). Players were classified using a k-means cluster analysis into four different age groups: 17–23 years, 24–27 years, 28–30 years, and 31–38 years. Linear mixed models were adjusted to compare the players’ match performance according to their age group and playing position (central defenders, external defenders, central midfielders, external midfielders, and forwards). The results showed that players aged between 31–38 years performed a significantly less total number of accelerations (ES=0.30–0.48) and decelerations (ES=0.29–0.49) in comparison with younger players. These age-related physical performance declines were more pronounced among central defenders, central midfielders, and forwards. However, no significant effects were obtained for players’ maximum acceleration and deceleration capacities. The current findings provide useful information for coaches and strength and conditioning specialists to better understand the effects of age on players’ physical performance and to develop age-tailored training programs.

Multi-level feature extraction network for person re-identification

In the task of Person re-identification (reID), the range of motion of pedestrians often spans multiple camera areas, and their motion direction and behavior cannot be constrained, and irrelevant people or objects in different scenes will also obtain target pedestrian information for us Cause interference. At the same time, the surveillance system also has many characteristics such as a fixed shooting angle of a single camera, different angles between different cameras, and low image resolution. These characteristics make the task of Person re-identification difficult. This paper proposes a Multi-level Feature Extraction Network (MFEN) based on SEResNet-50. Extracting richer and more diverse pedestrian features from poor-quality images will effectively improve the re-identification ability of the network, and MFEN can obtain Multistage key features in the image through the Feature Re-extraction Method (FRM) proposed in this paper. Experiments show that compared with AANet-50, MFEN has 3.85% /0.71% improvements of mAP/ Rank-1 on the Market1501 dataset, and 2.74% /1.28% improvements of mAP/ Rank-1 on the DukeMTMC-reID dataset.

Multiple Camera Fluorescence Detection for Real-Time PCR

The general polymerase chain reaction (PCR) amplifies DNA and analyzes the amplification results of the quantified DNA. Recently, real-time PCR has been developed to detect DNA amplification in various ways. The conventional camera-based system is too expensive and difficult to reduce device size. In this paper, we propose a low-cost, compact fluorescence detection system for real-time PCR systems using an open platform camera. To simplify the optics, four low-cost small cameras were fixedly placed, and the entire tube was divided into four quadrants to minimize the field of view. In addition, an effective image processing method was used to compensate. The proposed system measured the fluorescence detection performance on the basis of the amount of DNA using various fluorescent substances.