Alec G. Moore

Introduction to Machine Learning in Unity with Barracuda

Tue, 15 Feb 2022 18:54:00 +0000

Since I couldn’t find a lot of information on the web on how to get started with Barracuda in Unity, I thought I would go ahead and make a tutorial that covers the basics of getting started with the hopes that people could use it as a quick start guide before getting into more in-depth models.

The github for this project can be found here!

The Importance of Intersection Disambiguation for Virtual Hand Techniques

Tue, 21 Jul 2020 12:24:30 +0000

Beyond just finding valuable results, it was awesome getting to mentor two undergraduates and two high school students and help them to get an early academic publication and experience with Virtual Reality research! As always, abstract follows:

Some of the most widely used selection techniques for extended reality (XR) are based on virtual hand interactions. Many existing XR frameworks provide this functionality by default; however, their implementation can differ in slight, but important ways. When preparing to make a selection with a virtual hand technique, a user’s desired selection can potentially be ambiguous due to multiple intersections. Systems with varying underlying virtual hand implementations may yield contrasting selections due to resolving multiple intersections differently. This is particularly an issue when objects are smaller in size than the virtual hand representation and in dense environments. To demonstrate the importance of these differences, we present a virtual hand selection study comparing three methods that are currently used in popular XR frameworks for disambiguating selections: Closest Intersected, First Intersected, and Last Intersected. The results of our study show that the Closest Intersected method affords significantly faster selections, significantly fewer incorrect and missed selections, and yields significantly better effective throughput than the other two methods. These results show that using a framework’s built-in selection technique can significantly affect an XR application’s usability.

Check out the paper for more info!

VOTE: A Ray-casting Study of Vote Oriented Technique Enhancements

Sun, 19 Jul 2020 18:09:30 +0000

When making selections within 3D user interfaces (3DUIs), a user can fail to select a desired target despite indicating that target during most of the interaction process. This is due to numerous factors that can negatively impact which object is being indicated during the final confirmation step. In this paper, we present a novel vote-oriented technique enhancement (VOTE) for 3D selection that votes for indicated object each interaction frame and then selects the object with the most votes during confirmation. VOTE can be applied to nearly any 3D selection technique, as it does not require additional user input and does not require any prior knowledge of the environment or task. To demonstrate the effectiveness of VOTE, we present a ray-casting selection study that compared traditional, Snap-To, and VOTE ray-casting techniques for a standard multidirectional selection task. The results of our study show that VOTE afforded faster selections than traditional ray-casting and resulted in fewer incorrect selections than the Snap-To enhancement. Additionally, VOTE yielded significantly better effective throughput than traditional ray-casting and the Snap-To enhancement for selections within clustered environments.

Check out the paper for more info!

VUME: The Voluntary-Use Methodology for Evaluations

Tue, 31 May 2016 14:32:00 +0000

In an attempt to better understand how controlled research results impact actual voluntary use of 3D user interfaces (3D Uls), we developed a new evaluation approach. Using this approach, we conducted two studies evaluating two head-mounted displays (HMDs) - a Sensics zSight and an Oculus Rift Development Kit 1 (DK1). The results of the first study indicate that the DK1 affords significantly better user performances. In the second study, we used a between-subjects design to determine if participants would voluntarily explore and interact with a virtual environment more with the DK1 than the zSight. We did not find a significant difference between the two HMDs, but statistically proved that the HMDs were equivalent. This indicates that results found in controlled evaluations do not always play a significant role in the voluntary use of a 3D UI.

Check out the paper for more info!

A reproducible olfactory display for exploring olfaction in immersive media experiences

Thu, 22 Oct 2015 19:59:00 +0000

Olfaction, the sense of smell, is an important perceptual function. It has been shown to enhance the quality of life and to facilitate memory recall. However, despite its importance and potential benefits, olfaction is often neglected when creating immersive media experiences (IMEs). The probable reason is that olfactory displays are not readily available. We have developed a reproducible olfactory display that is simple and affordable to make olfactory displays more available. More importantly, we have conducted three exploratory studies on the effects of olfaction. The results of these studies indicate that olfaction is a nuanced sense that can have both positive and negative effects on user experiences. Hence, further research is needed to better understand and design for olfaction in IMEs.

Check out the paper for more info!

Using Remote Control Aerial Vehicles to Study Variability of Airborne Particulates

Sat, 05 Sep 2015 19:46:00 +0000

Airborne particulates play a significant role in the atmospheric radiative balance and impact human health. To characterize this impact, global-scale observations and data products are needed. Satellite products allow for this global coverage but require in situ validations. This study used a remote-controlled aerial vehicle to look at the horizontal, vertical, and temporal variability of airborne particulates within the first 150 m of the atmosphere. Four flights were conducted on December 4, 2014, between 12:00 pm and 5:00 pm local time. The first three flights flew a pattern of increasing altitude up to 140 m. The fourth flight was conducted at a near-constant altitude of 60 m. The mean PM2.5 concentration for the three flights with varying altitude was 36.3 μg/m³, with the highest concentration occurring below 10 m altitude. The overall vertical variation was very small with a standard deviation of only 3.6 μg/m³. PM2.5 concentration also did not change much throughout the day with mean concentrations for the altitude-varying flights of 35.1, 37.2, and 36.8 μg/m³. The fourth flight, flown at a near-constant altitude, had a lower concentration of 23.5 μg/m³.

Check out the paper for more info!

Estimation and Bias Correction of Aerosol Abundance using Data-Driven Machine Learning and Remote Sensing

Fri, 14 Aug 2015 15:30:39 +0000

Air quality information is increasingly becoming a public health concern, since some of the aerosol particles pose harmful effects to peoples health. One widely available metric of aerosol abundance is the aerosol optical depth (AOD). The AOD is the integrated light extinction coefficient over a vertical atmospheric column of unit cross section, which represents the extent to which the aerosols in that vertical profile prevent the transmission of light by absorption or scattering. The comparison between the AOD measured from the ground-based Aerosol Robotic Network (AERONET) system and the satellite MODIS instruments at 550 nm shows that there is a bias between the two data products. We performed a comprehensive search exploring possible factors which may be contributing to the inter-instrumental bias between MODIS-Aqua land data set and AERONET. The analysis used several measured variables, including the MODIS AOD, as input in order to train a neural network in regression mode to predict the AERONET AOD values. This not only allowed us to obtain an estimate, but also allowed us to infer the optimal sets of variables that played an important role in the prediction. In addition, we applied machine learning to infer the global abundance of ground level PM2.5 from the AOD data and other ancillary satellite and meteorology products. This research is part of our goal to provide air quality information, which can also be useful for global epidemiology studies.

Check out the paper for more info!

The Effects of Olfaction on Training Transfer for an Assembly Task

Fri, 14 Aug 2015 15:26:39 +0000

Context-dependent memory studies have indicated that olfaction, the sense of smell, has a special odor memory that can significantly improve recall in some cases. Virtual reality (VR), which has been investigated as a training tool, could feasibly benefit from odor memory by incorporating olfactory stimuli. There have been a few studies on this concept for semantic learning, but not for procedural training. To address this gap in knowledge, we investigated the effects of olfaction on the transfer of knowledge from training to next-day execution for building a complex LEGO jet-plane model. Our results indicate that the pleasantness of an odor significantly affects training transfer more than whether the encoding and recall contexts match.

Check out the paper for more info!

Wedge: A Musical Interface for Building and Playing Composition-Appropriate Immersive Environments

Thu, 13 Aug 2015 21:58:39 +0000

We present Wedge - a novel musical interface designed for building an immersive environment of notes that is appropriate for the musician and the composition being played. The interface uses two simple gestures for building and playing composition appropriate immersive environments. A horizontal-finger gesture allows users to place note “wedges” from a virtual keyboard and to move them between snapping grids with their own local coordinate systems. A vertical-finger gesture allows users to play the notes and chords by striking the wedges like strings, with the velocity determining the volume of the sounds.

Check out the paper for more info!