Neural networks have become increasingly adopted in society over the last few years. As neural networks consume a lot of energy to train, reducing the energy consumption of these networks is desirable from an environmental perspective. Spiking neural network is a type of neural network inspired by the human brain which is significantly more energy efficient than traditional neural networks. However, there is little research about how the hyper parameters of these networks affect the relationship between accuracy and energy. The aim of this report is therefore to analyse this relationship. To do this, we measure the energy usage of training several different spiking network models. The results of this study shows that the choice of hyper-parameters in a neural network does affect the efficiency of the network. While correlation between any individual factors and energy consumption is inconclusive, this work could be used as a springboard for further research in this area. / Under de senaste åren har neuronnät blivit allt vanligare i samhället. Eftersom neuronnät förbrukar mycket energi för att träna dem är det önskvärt ur miljösynpunkt att minska energiförbrukningen för dessa nätverk. Spikande neuronnät är en typ av neuronnät inspirerade av den mänskliga hjärnan som är betydligt mer energieffektivt än traditionella neuronnät. Det finns dock lite forskning om hur hyperparametrarna i dessa nätverk påverkar sambandet mellan noggrannhet och energi. Syftet med denna rapport är därför att analysera detta samband. För att göra detta mäter vi energiförbrukningen vid träning av flera olika modeller av spikande neuronnät-modeller. Resultaten av denna studie visar att valet av hyperparametrar i ett neuronnät påverkar nätverkets effektivitet. Även om korrelationen mellan enskilda faktorer och energiförbrukning inte är entydig kan detta arbete användas som en startpunkt för ytterligare forskning inom detta område.

Event-based vision offers high dynamic range, time resolution and lower latency than conventional frame-based vision sensors. These attributes are useful in varying light condition and fast motion. However, there are no neural network models and training protocols optimized for object detection with event data, and conventional artificial neural networks for frame-based data are not directly suitable for that task. Spiking neural networks are natural candidates but further work is required to develop an efficient object detection architecture and end-to-end training protocol. For example, object detection in varying light conditions is identified as a challenging problem for the automation of construction equipment such as earth-moving machines, aiming to increase the safety of operators and make repetitive processes less tedious. This work focuses on the development and evaluation of a neural network for object detection with data from an event-based sensor. Furthermore, the strengths and weaknesses of an event-based vision solution are discussed in relation to the known challenges described in former works on automation of earth-moving machines. A solution for object detection with event data is implemented as a modified YOLOv3 network with spiking convolutional layers trained with a backpropagation algorithm adapted for spiking neural networks. The performance is evaluated on the N-Caltech101 dataset with classes for airplanes and motorbikes, resulting in a mAP of 95.8% for the combined network and 98.8% for the original YOLOv3 network with the same architecture. The solution is investigated as a proof of concept and suggestions for further work is described based on a recurrent spiking neural network.

<div>Event-Based cameras, also known as neuromophic cameras or dynamic vision sensors, are an imaging modality that attempt to mimic human eyes by asynchronously measuring contrast over time. If the contrast changes sufficiently then a 1-bit event is output, indicating whether the contrast has gone up or down. This stream of events is sparse, and its asynchronous nature allows the pixels to have a high dynamic range and high temporal resolution. However, these events do not encode the intensity of the scene, resulting in an inverse problem to estimate intensity images from the event stream. Hybrid event-based cameras, such as the DAVIS camera, provide a reference intensity image that can be leveraged when estimating the intensity at each pixel during an event. Normally, inverse problems are solved by formulating a forward and prior model and minimizing the associated cost, however, for this problem, the Plug and Play (P&P) algorithm is used to solve the inverse problem. In this case, P&P replaces the prior model subproblem with a denoiser, making the algorithm modular, easier to implement. We propose an idealized forward model that assumes the contrast steps measured by the DAVIS camera are uniform in size to simplify the problem. We show that the algorithm can swiftly reconstruct the scene intensity at a user-specified frame rate, depending on the chosen denoiser’s computational complexity and the selected frame rate.</div>

Image classification using data from neuromorphic vision sensors is a challenging task that affects the use of dynamic vision sensor cameras in real- world environments. One impeding factor is noise in the neuromorphic event stream, which is often generated by the dynamic vision sensors themselves. This means that effective noise filtration is key to successful use of event- based data streams in real-world applications. In this paper we harness two feature representations of neuromorphic vision data in order to apply conventional frame-based image tools on the neuromorphic event stream. We use a standard noise filter to evaluate the effectiveness of noise filtration using a popular dataset converted to neuromorphic vision data. The two feature representations are the best-of-class standard Histograms of Averaged Time Surfaces (HATS) and a simpler grid matrix representation. To evaluate the effectiveness of the noise filter, we compare classification accuracies using various noise filter windows at different noise levels by adding additional artificially generated Gaussian noise to the dataset. Our performance metrics are reported as classification accuracy. Our results show that the classification accuracy using frames generated with HATS is not significantly improved by a noise filter. However, the classification accuracy of the frames generated with the more traditional grid representation is improved. These results can be refined and tuned for other datasets and may eventually contribute to on- the- fly noise reduction in neuromorphic vision sensors. / Händelsekameror är en ny typ av kamera som registrerar små ljusförändringar i kamerans synfält. Sensorn som kameran bygger på är modellerad efter näthinnan som finns i våra ögon. Näthinnan är uppbyggd av tunna lager av celler som omvandlar ljus till nervsignaler. Eftersom synsensorer efterliknar nervsystemet har de getts namnet neuromorfiska synsensorer. För att registrera små ljusförändringar måste dessa sensorer vara väldigt känsliga vilket även genererar ett elektroniskt brus. Detta brus försämrar kvalitén på signalen vilket blir en förhindrande faktor när dessa synsensorer ska användas i praktiken och ställer stora krav på att hitta effektiva metoder för brusredusering. Denna avhandling undersöker två typer av digitala framställningar som omvandlar signalen ifrån händelsekameror till något som efterliknar vanliga bilder som kan användas med traditionella metoder för bildigenkänning. Vi undersöker brusreduseringens inverkan på den övergripande noggrannhet som uppnås av en artificiell intelligens vid bildigenkänning. För att utmana AIn har vi tillfört ytterligare normalfördelat brus i signalen. De digitala framställningar som används är dels histogram av genomsnittliga tidsytor (eng. histograms of averaged time surfaces) och en matrisrepresentation. Vi visar att HATS är robust och klarar av att generera digitala framställningar som tillåter AIn att bibehålla god noggrannhet även vid höga nivåer av brus, vilket medför att brusreduseringens inverkan var försumbar. Matrisrepresentationen gynnas av brusredusering vid högre nivåer av brus.