Music has been heavily dependent on computers since the 80’s, and with each new advancement in computers, new advancements in music soon follow. I asked myself how will the future of pattern recognition affect the music industry? This paper explores how complex pattern recognition systems, specifically artificial neural networks, might affect the next advancements in music. Melodies from hundreds of popular songs from multiple genres were analyzed by a neural network. Everything from "Tiny Dancer" by Elton John to "I’m a Barbie girl" by Aqua were used to train the network. By having the neural network produce midi tracks, I was able to import them into a professional music producing program, named Logic Pro. The neural network’s songs were then produced as if they were generated by an artist paying for studio time. The final song was uplifting, melodic and entertaining. This demonstrates the effectiveness of neural networks as a creative musical tool. The model I propose is an improvement on an existing model, an encoder to decoder neural network written in TensorFlow. Because of music’s heavy dependency on previous events LSTM(Long Short Term Memory) cells were used. To improve preprocessing I considered transposing all of the songs to the same key, C-major, this restricts the model to track less notes by limiting the scale, while still maintaining the melody of the song. I found that working with a batch of short midi loops generated by the network worked best for creative output.

With the spread of large digital painting collections, the demand for automatic theme classification has risen in both academic and commercial fields. Many researchers have investigated various techniques to implement automatic painting classification and developed several approaches to improve classification accuracy. SURF is one of the most commonly used local invariant feature descriptors for painting classification. However, the common method of classification based on traditional SURF local feature description makes the color feature information not comprehensive. Also, the feature space is high dimensional and sparse which will result in low classification accuracy, data redundancy and a time-consuming process. To overcome this drawback, we suggest a new method of classifying painting styles with bag of words (BOW) model based on the RGB-SURF algorithm. First, split RGB channels into separate images for each painting. Then, independently calculate key points and descriptor of each RGB channel using SURF and combine descriptors of each channel to obtain overall RGB-SURF feature for each painting. This approach is to compensate the loss of color feature brought by traditional SURF descriptor which mainly treats colorful images as gray images. Next, construct the visual vocabulary by reducing the number of features through quantization of feature space using K-means clustering to obtain BOW model. The histogram described by appearance frequency of the visual words is used to represent the content of the image. As a result, every image is viewed as a bag full of visual words. Finally, establish the support vector machine (SVM) classifier based on radial basis function (RBF) with the data above for training and testing. Compared to classification method based on the traditional SURF feature descriptor and other painting classification approaches, BOW method based on RGB-SURF algorithm shows the highest average classification accuracy, which means this method produces the most consistent automatic classification results with manual classification.

Though many teleoperation interfaces have been made for mobile manipulation tasks, the majority of these interfaces are inaccessible to people with severe physical disabilities, and the interfaces that are accessible are extremely slow, requiring around three minutes to complete a simple manipulation task, such as picking up and moving a soda can a few feet. People with severe physical disabilities are those who could benefit the most from these teleoperation interfaces, because through teleoperation, they can complete tasks that they can not perform on their own, such as opening the door, fetching an object, or making the bed. For a teleoperation interface to be accessible to people with severe physical disabilities, it must rely on cursor movement and clicks, or speech commands. Because the current accessible interfaces are slow, my research goal is to design new interfaces that are both accessible and brisk. I have designed three of these interfaces so far and am in the process of finishing implementation. The interfaces use two orthogonal views instead of the standard one view, thus allowing the user to grasp a complete picture of the robot and the world without needing to change the viewing angle, a slow and frustrating process. The three interfaces which allow for control of the end effector consist of two types of voice commands, and one button interface. Soon I will be conducting user studies with the new interfaces to quantify their improvement over the current standard.

Dementia is a neurological condition that impairs memory, cognitive functions, and the performance of daily activities. It is estimated that 5.5 million Americans suffered from dementia in 2017. Currently, the predominant method used to treat persons with dementia is through pharmacological means. However, through a review of current research we have found evidence to suggest that cognitive behavior therapies (CBT) may be more beneficial to patients, with the potential to reverse some symptoms of dementia. Our study aims to improve CBT through virtual environments. Virtual Reality (VR) is not merely a visual platform but an immersive environment that stimulates the brain in the same way real environments do. Patients with dementia living nursing homes are rarely exposed to different environmental interactions or thought inducing stimuli, resulting in an escalated progression of their condition and a loss of “sense of self.” CBT, specifically reminiscence therapy, has shown promising results in improving cognitive functioning as well as morale. We hypothesize that if we use visual and audio equipment as well as tangible artifacts, we could create an immersive therapy to stimulate cognitive functions. We track certain markers like head and eye movement to track the progression of the disease as well. In collaboration with a UK based game development company we have identified a software with both therapeutic and a diagnostic capabilities. We have made connections with senior centers in the area to identify persons wishing to participate VR therapy sessions. This project will offer understanding of how implementing VR therapy can improve patients’ morale and cognitive capabilities, as well as exploring the use of VR as a diagnostic tool.

Augmented Reality, the superimposition of digital features into the real world environment, has already made great strides in fields such as entertainment and advertising. However, the use of compound technology that bridges reality and the digital world has yet to be implemented cogently and thoroughly into the manufacturing field. This first part of this research paper explores the range of augmented reality hardware and software interfaces in fields such as medicine, academics, entertainment and advertising through a review of other scientific studies on augmented reality. Next, this research applies these observations to the current preliminary augmented reality trends specifically in the manufacturing field. Taking into account the positive and negative cognitive effects of augmented reality systems on attention, motivation, and ease of use, this paper then further delves into the potential uses of augmented reality in manufacturing. In the next section, this paper uses its examination of augmented reality’s ability to cater the manufacturing community’s needs to evaluate the advantages of implementing augmented reality technology in conjunction with human ability to improve the manufacturing process. Finally, the last section of this paper synthesizes the ideas presented into a set of principles for augmented reality use in manufacturing, and gives real-world examples of these principles’ efficient applications.

Augmented reality (AR), which overlays virtual content on top of the user's perception of the real world, is now beginning to enter the consumer market. Besides smartphone platforms, early-stage head-mounted displays such as the Microsoft HoloLens are being actively developed. Many of the most compelling uses of these AR technologies are multi-user: for instance, in-person collaborative tools, multiplayer gaming, and telepresence. Although AR technologies enable new forms of interaction, new security and privacy challenges will also arise when users can augment each other's reality, and it is imperative that these challenges be addressed while the technology is still new and highly malleable. In this work, I explore these emerging challenges in secure and private content sharing for multi-user AR. I systematize design goals for security and functionality that an AR content sharing framework should support, and I design and prototype a framework for the HoloLens that meets these goals. By evaluating my framework against representative application case studies, I show that it meets desired security and functionality goals flexibly across a range of use cases. Preliminary investigations into developer effort suggest that applications' content sharing needs can be achieved in relatively few lines of code. I plan to convert my research prototype into an open-source toolkit so developers can address these challenges in practice. This work opens up directions for future research in how these underlying paradigms should manifest to users in the form of an application's user interface. By building foundations for secure multi-user AR content sharing, my work takes steps toward allowing AR to securely reach its full potential.

This research attempt proposes a novel approach to processing and utilizing aerial imagery and data for mapping. With the advancement of unmanned aerial systems (UAS), airborne data has gained considerable momentum in recent decades, revolutionizing methodologies of geographic research, law enforcement, agriculture, and mapping. Specifically, advances in consumer electronics significantly eased access to UAS and development in parallel computing has enabled machine learning on large data sets. This paper proposes an alternative method to the current practice of processing airborne data sets. Instead of obtaining aerial imagery of an object and merely presenting those objects as pixels on a map, the proposed method uses machine learning techniques to recognize the object, assign parameters to the object and render the object in a fashion that is most efficient and understandable to the end user. The object is no longer merely a set of pixels but a list of classified objects with associated parameters, an abstraction analogous to cognitive processes. This approach allows a single data set to be used in multiple fashions. For example, the data can be used to simulate changes in the environment or render the environment in different scenarios. This paper explores the methodology for collecting airborne data sets for this process and presents several cases of this using flight test data.

Small biologically inspired robots have difficulty landing after flight. This work is intended to explore a possible method for small robots in flight to attach to an overhanging surface. The goal of this work is to create an insect sized robot that is capable of jumping from a horizontal surface to an inverted surface and mechanically grip its destination. Previous work on attaching to a surface from flight have used electrostatic or electromagnetic copper coils or adhesive. My method for attaching to a surface was inspired by the sharp-hooked feet of beetles and uses hooks the same size as ten stripe beetle’s to hold the surface. The advantage of this attachment method is that it does not need to be electrically powered to maintain a hold on the surface and is lighter weight than adhesives or copper coils used in other methods. I used an iterative design process to build and test different shapes and angles of hooks as well as to develop a jumping robot base structure that could support the grabbing arms and quickly repeat jump tests. I captured photos and video of beetle and grasshopper legs in motion and recreated the hooks they have on their legs as a starting point for the hooks I developed. In the future, this mechanism could be attached to robots capable of controlled flight, allowing them to perch on a surface indefinitely and then release the surface to resume flight.