Seventy percent of stroke survivors experience loss of upper limb function after stroke and rehabilitative therapy is the only option to reduce impairments. However, uncertainty remains as to the optimal dose of therapy that should be prescribed. It has been suggested to report multiple parameters of dose, to increase standardization within the field, and to gain a better understanding of the dose-response relationship. This study investigated the automatic quantification of multiple dose parameters in a rat model of ischemic stroke, with rehabilitation paradigms whereby rats repeatedly grasp for food pellets to train their forelimb function. Starting 7 days post-stroke, groups of rats received 4, 8, or 12 rehabilitative training sessions for 10 days, practicing either high-quality (precision practice) or less skilled (mass practice) reaching movements. Pellet consumption was measured after each session and various metrics were analyzed using deep learning-based software (DeepLabCut, DLC) to represent parameters of dose intensity (number of reaches, paw path length) and session density (time on task). Functional outcome was assessed with the Montoya staircase task. Computer algorithms were validated against human analysis, demonstrating reach detection accuracy and reliability >80%. Interestingly, the number of training sessions did not alter the accumulated movement practice across rehabilitation, in either task. However, the number of sessions inversely affected training intensity, resulting in more forelimb use per session in rats with 4 sessions compared to 12 sessions. We found strong positive correlations between the number of reaches, time on task, paw path length, and pellets consumed in the precision practice, but only between reaches and pellets consumed in mass practice. This work demonstrates the quantification of multiple dose parameters using deep learning software and shows subtle differences between the two commonly used forelimb training tasks. Moreover, our data suggest that rehabilitative training at a frequency that is too high may negatively impact performance per session.

Pose estimation uses computer vision to predict how a body moves. The likeliness off different movements is predicted with a neural network and the most likely pose is predicted. With DeepLabCut, an open source software package for 3D animal pose estimation, information about animals behaviour and movement can be extracted. In this report the pose estimation of horses four hooves is used. By looking at the position of the hooves different gaits can be identified. Horses used for riding in the major disciplines in Sweden have three different gaits, walk, trot and canter. Walk is a four-stoke gait, trot is two-stoke and canter is three-stoke. This can be used to classify the different gaits. By looking at the hooves movement in vertical position over time and fitting a sinewave to the data it is possible to see the phase difference in the hooves movement. For walk and trot there was a significant pattern which was easy to identify and corresponded well to the theory of horses movement. For canter our pre-trained model lacked in accuracy, so the output data were insufficient. Therefore it was not possible to find a significant pattern for canter which corresponds to the theory of horses movements. The Fourier Transform were also tested to classify the gaits and when plotted it was possible to detect the different gaits, but not significant enough to be reliable for different horses in different sizes running in different paces. It was also possible to add the data for all four hooves together and fit a sinewave to the added data, and then compare it with the sinewaves for each hoof separately. Depending on the gait the frequency of the sinewaves differed between the hooves separately and added together and the gaits could be identified.

This thesis investigates whether Computer Vision can be a useful tool in interpreting the behaviors of monitored horses. In recent years, research in the field of Computer Vision has primarily focused on people, where pose estimation and action recognition are popular research areas. The thesis presents a pose classification network, where input features are described by estimated 2D key- points of horse body parts. The network output classifies three poses: ’Head above the wither’, ’Head aligned with the wither’ and ’Head below the wither’. The 2D reconstructions of keypoints are obtained using DeepLabCut applied to raw video surveillance data of a single horse. The estimated keypoints are then fed into a Multi-layer preceptron, which is trained to classify the mentioned classes. The network shows promising results with good performance. We found label noise when we spot-checked random samples of predicted poses and comparing them to the ground truth, as some of the labeled data consisted of false ground truth samples. Despite this fact, the conclusion is that satisfactory results are achieved with our method. Particularly, the keypoint estimates were sufficient enough for these poses for the model to succeed to classify a hold-out set of poses. / Uppsatsen undersöker främst om datorseende kan vara ett användbart verktyg för att tolka beteendet hos övervakade hästar. Under de senaste åren har forskning inom datorseende främst fokuserat på människor, där pose-estimering och händelseigenkänning är populära forskningsområden. Denna avhandling presenterar ett poseklassificeringsnätverk där indata beskrivs av uppskattade 2Dnyckelpunkter (eller så kallade intressepunkter) för hästkroppsdelar. Nätverket klassificerar tre poser: ’Huvud ovanför manken’, ’Huvud i linje med manken’ och ’Huvudet nedanför manken’. 2D-rekonstruktioner av nyckelpunkter erhålls med hjälp av DeepLabCut, applicerad på rå videoövervakningsdata för en häst. De uppskattade nyckelpunkterna matas sedan in i ett flerskikts- preceptron, som tränas för att klassificera de nämnda klasserna. Nätverket visar lovande resultat med bra prestanda. Vi hittade brus i etiketterna vid slumpmässiga stickprover av förutspådda poser som jämfördes med sanna etiketter där några etiketter bestod av falska sanna etiketter. Trots detta är slutsatsen att tillfredsställande resultat uppnås med vår metod. Speciellt var de estimerade nyckelpunkterna tillräckliga för dessa poser för att nätverket skulle lyckas med att klassificera ett separat dataset av samma osedda poser.

Deer Vehicle Collisions (DVCs) are a global problem that is not only resulting in seriousinjuries to humans but also results in loss of human and deer lives. Deer are more active and less attentive during the mating and hunting seasons. Roadside deer activity such as feeding and strolling along the roadside has a significant correlation with DVCs. To mitigate DVCs, several strategies were used that include vegetation management, fences, underpasses and overpasses, population reduction, warning signs and animal detection systems (ADS). These strategies vary in their efficacy. These strategies may help to reduce DVCs. However, they are not always easily feasible due to false alarms, high cost, unsuitable terrain, land ownership, and other factors. Thus, DVCs are increasing due to the increase in number of vehicles and the absence of intelligent highway safety and alert systems. Detecting deer in real-time on our roads is a challenging problem. Thus, this research work proposed a deer detection and movement DDM technique to automate DVCs mitigation system. The DDM combines computer vision, artificial intelligent methods with deep learning techniques. DDM includes two main deep learning algorithms 1)onestage deep learning algorithm based on Yolov5 that generates a detection model(DM) to detect deer and 2) deep learning algorithm developed by python toolkit DeepLabCut to generate movement model(MM) for detecting the movement of the deer. The proposed method can detect deer with 99.7% precision and succeeds to ascertain if the deer is moving or static with an inference speed of 0.29s. The proposed method can detect deer with 99.7% precision and using DeepLabCut toolkit on the detected deer we can ascertain if the deer is moving or static with an inference speed of 0.29s.

No description available.