Many video feature extraction algorithms have been purposed, such as STIP, HOG3D, and Dense Trajectories. These algorithms are often referred to as “handcrafted” features as they were deliberately designed based on some reasonable considerations. However, these algorithms may fail when dealing with high-level tasks or complex scene videos. Due to the success of using deep convolution neural networks (CNNs) to extract global representations for static images, researchers have been using similar techniques to tackle video contents. Typical techniques first extract spatial features by processing raw images using deep convolution architectures designed for static image classifications. Then simple average, concatenation or classifier-based fusion/pooling methods are applied to the extracted features. I argue that features extracted in such ways do not acquire enough representative information since videos, unlike images, should be characterized as a temporal sequence of semantically coherent visual contents and thus need to be represented in a manner considering both semantic and spatio-temporal information.
In this thesis, I propose a novel architecture to learn semantic spatio-temporal embedding for videos to support high-level video analysis. The proposed method encodes video spatial and temporal information separately by employing a deep architecture consisting of two channels of convolutional neural networks (capturing appearance and local motion) followed by their corresponding Fully Connected Gated Recurrent Unit (FC-GRU) encoders for capturing longer-term temporal structure of the CNN features. The resultant spatio-temporal representation (a vector) is used to learn a mapping via a Fully Connected Multilayer Perceptron (FC-MLP) to the word2vec semantic embedding space, leading to a semantic interpretation of the video vector that supports high-level analysis. I evaluate the usefulness and effectiveness of this new video representation by conducting experiments on action recognition, zero-shot video classification, and semantic video retrieval (word-to-video) retrieval, using the UCF101 action recognition dataset.
The dissertation utilizes Interval Data (ID) and establishes three different frameworks to identify electricity losses, predict electricity consumption and detect anomalies using data mining, deep learning, and mathematical models. The process of energy analytics integrates with the computational science and contributes to several objectives which are to
1. Develop a framework to identify both technical and non-technical losses using clustering and semi-supervised learning techniques.
2. Develop an integrated framework to predict electricity consumption using wavelet based data transformation model and deep learning algorithms.
3. Develop a framework to detect anomalies using ensemble empirical mode decomposition and isolation forest algorithms.
With a thorough research background, the first phase details on performing data analytics on the demand-supply database to determine the potential energy loss reduction potentials. Data preprocessing and electricity prediction framework in the second phase integrates mathematical models and deep learning algorithms to accurately predict consumption. The third phase employs data decomposition model and data mining techniques to detect the anomalies of institutional buildings.
sampling for both spatial and angular dimensions. Single-shot light field cameras
sacrifice spatial resolution to sample angular viewpoints, typically by multiplexing
incoming rays onto a 2D sensor array. While this resolution can be recovered using
compressive sensing, these iterative solutions are slow in processing a light field. We
present a deep learning approach using a new, two branch network architecture,
consisting jointly of an autoencoder and a 4D CNN, to recover a high resolution
4D light field from a single coded 2D image. This network decreases reconstruction
time significantly while achieving average PSNR values of 26-32 dB on a variety of
light fields. In particular, reconstruction time is decreased from 35 minutes to 6.7
minutes as compared to the dictionary method for equivalent visual quality. These
reconstructions are performed at small sampling/compression ratios as low as 8%,
allowing for cheaper coded light field cameras. We test our network reconstructions
on synthetic light fields, simulated coded measurements of real light fields captured
from a Lytro Illum camera, and real coded images from a custom CMOS diffractive
light field camera. The combination of compressive light field capture with deep
learning allows the potential for real-time light field video acquisition systems in the
future.
This thesis develops a classification method to investigate the performance of FDG-PET as an effective biomarker for Alzheimer's clinical group classification. This involves dimensionality reduction using Probabilistic Principal Component Analysis on max-pooled data and mean-pooled data, followed by a Multilayer Feed Forward Neural Network which performs binary classification. Max pooled features result into better classification performance compared to results on mean pooled features. Additionally, experiments are done to investigate if the addition of important demographic features such as Functional Activities Questionnaire(FAQ), gene information helps improve performance. Classification results indicate that our designed classifiers achieve competitive results, and better with the additional of demographic features.
The dissertation outlines novel domain adaptation approaches across different feature spaces; (i) a linear Support Vector Machine model for domain alignment; (ii) a nonlinear kernel based approach that embeds domain-aligned data for enhanced classification; (iii) a hierarchical model implemented using deep learning, that estimates domain-aligned hash values for the source and target data, and (iv) a proposal for a feature selection technique to reduce cross-domain disparity. These adaptation procedures are tested and validated across a range of computer vision applications like object classification, facial expression recognition, digit recognition, and activity recognition. The dissertation also provides a unique perspective of domain adaptation literature from the point-of-view of linear, nonlinear and hierarchical feature spaces. The dissertation concludes with a discussion on the future directions for research that highlight the role of domain adaptation in an era of rapid advancements in artificial intelligence.
Recovery of CS video frames is far more complex than still images, which are known to be (approximately) sparse in a linear basis such as the discrete cosine transform. By combining sparsity of individual frames with an optical flow-based model of inter-frame dependence, the perceptual quality and peak signal to noise ratio (PSNR) of reconstructed frames is improved. The efficacy of this approach is demonstrated for the cases of \textit{a priori} known image motion and unknown but constant image-wide motion.
Although video sequences can be reconstructed from CS measurements, the process is computationally costly. In autonomous systems, this reconstruction step is unnecessary if higher-level conclusions can be drawn directly from the CS data. A tracking algorithm is described and evaluated which can hold target vehicles at very high levels of compression where reconstruction of video frames fails. The algorithm performs tracking by detection using a particle filter with likelihood given by a maximum average correlation height (MACH) target template model.
Motivated by possible improvements over the MACH filter-based likelihood estimation of the tracking algorithm, the application of deep learning models to detection and classification of compressively sensed images is explored. In tests, a Deep Boltzmann Machine trained on CS measurements outperforms a naive reconstruct-first approach.
Taken together, progress in these three areas of CS inference has the potential to lower system cost and improve performance, opening up new applications of CS video cameras.
The purpose of this project is to create a useful tool for musicians that utilizes the harmonic content of their playing to recommend new, relevant chords to play. This is done by training various Long Short-Term Memory (LSTM) Recurrent Neural Networks (RNNs) on the lead sheets of 100 different jazz standards. A total of 200 unique datasets were produced and tested, resulting in the prediction of nearly 51 million chords. A note-prediction accuracy of 82.1% and a chord-prediction accuracy of 34.5% were achieved across all datasets. Methods of data representation that were rooted in valid music theory frameworks were found to increase the efficacy of harmonic prediction by up to 6%. Optimal LSTM input sizes were also determined for each method of data representation.
Carbohydrate counting has been shown to improve HbA1c levels for people with diabetes. However, the learning curve and inconvenience of carbohydrate counting make it difficult for patients to adhere to it. A deep learning model is proposed to identify food from an image, where it can help the user manage their carbohydrate counting. This early model has a 68.3% accuracy of identifying 101 different food classes. A more refined model in future work could be deployed into a mobile application to identify food the user is about to consume and log it for easier carbohydrate counting.