Ross T. Sowell

Video Manipulation Techniques for the Protection of Privacy in Remote Presence Systems

Systems that give control of a mobile robot to a remote user raise privacy concerns about what the remote user can see and do through the robot. We aim to preserve some of that privacy by manipulating the video data that the remote user sees. Through two user studies, we explore the effectiveness of different video manipulation techniques at providing different types of privacy. We simultaneously examine task performance in the presence of privacy protection. In the first study, participants were asked to watch a video captured by a robot exploring an office environment and to complete a series of observational tasks under differing video manipulation conditions. Our results show that using manipulations of the video stream can lead to fewer privacy violations for different privacy types. Through a second user study, it was demonstrated that these privacy-protecting techniques were effective without diminishing the task performance of the remote user.

VolumeViewer: An Interactive Tool for Fitting Surfaces to Volume Data

Recent advances in surface reconstruction algorithms [BM07, LBD 08] allow surfaces to be built from contours lying on non-parallel planes. Such algorithms allow users to construct surfaces of similar quality more efficiently by using a small set of oblique contours, rather than many parallel contours. However, current medical imaging systems do not provide tools for sketching contours on oblique planes. In this paper, we take the first steps towards bridging the gap between the new surface reconstruction technologies and putting those methods to use in practice. We develop a novel interface for modeling surfaces from volume data by allowing the user to sketch contours on arbitrarily oriented cross-sections of the volume, and we examine the users’ ability to contour the same structures using oblique cross-sections with similar consistency as they can using parallel cross-sections. We measure the inter-observer and intra-observer variability of trained physicians contouring on oblique cross-sections of real patient data as compared to the traditional parallel cross-sections, and show that the variation is much higher for oblique contouring. We then show that this variability can be greatly reduced by integrating a collection of training images into the interface.

Using Video Manipulation to Protect Privacy in Remote Presence Systems

Remote presence systems that allow remote operators to physically move around the world, observe it, and, in some cases, manipulate it, introduce a new set of privacy concerns. Traditional telepresence systems allow remote users to passively observe, forcing them to look at whatever the camera is pointing at. If we want something to remain private, then we simply do not put it in front of the camera. Remote presence systems, on the other hand, allow active observation, and put the control of the camera in the hands of the remote operator. They can drive around, and look at the world from different viewpoints, which complicates privacy protection.

Privacy, Utility, and Cognitive Load in Remote Presence Systems

As teleoperated robot technology becomes cheaper, more powerful, and more reliable, remotely-operated telepresence robots will become more prevalent in homes and businesses, allowing visitors and business partners to be present without the need to travel. Hindering adoption is the issue of privacy: an Internet-connected telepresence robot has the ability to spy on its local area, either for the remote operator or a third party with access to the video data. Additionally, since the remote operator may move about and manipulate objects without local-user intervention, certain typical privacy-protecting techniques such as moving objects to a different room or putting them in a cabinet may prove insufficient. In this paper, we examine the effects of three whole-image filters on the remote operators ability to discern details while completing a navigation task.

A Method for Establishing Correspondences Between Hand-Drawn and Sensor-Generated Maps

Maps, and specifically floor plans, are useful for planning a variety of tasks from arranging furniture to designating conceptual or functional spaces (e.g., kitchen, walkway). However, maps generated directly from robot sensor data can be hard to interpret and use for this purpose, especially for individuals who are not used to them, because of sensor and odometry measurement errors and the probabilistic nature of the mapping algorithms themselves. In this paper, we present an algorithm for quickly laying a floor plan (or other conceptual map) onto a map generated from sensor data, creating a one-to-one mapping between the two This allows humans interacting with the robot to use a more readily-understandable representation of the world, while the robot itself uses the sensor-generated map.

Guided Structure-Aligned Segmentation of Volumetric Data

Segmentation of volumetric images is considered a time and resource intensive bottleneck in scientific endeavors. Automatic methods are becoming more reliable, but many data sets still require manual intervention. Key difficulties include navigating the 3D image, determining where to place marks, and maintaining consistency between marks and segmentations. Clinical practice often requires segmenting many different instances of a specific structure. In this research we leverage the similarity of a repeated segmentation task to address these difficulties and reduce the cognitive load for segmenting on non-traditional planes. We propose the idea of guided contouring protocols that provide guidance in the form of an automatic navigation path to arbitrary cross sections, example marks from similar data sets, and text instructions. We present a user study that shows the usability of this system with non-expert users in terms of segmentation accuracy, consistency, and efficiency.

Modeling surfaces from volume data using nonparallel contours

Magnetic resonance imaging (MRI) and computed tomography (CT) scanners have long been used to produce three-dimensional samplings of anatomy elements for use in medical visualization and analysis. From such datasets, physicians often need to construct surfaces representing anatomical shapes in order to conduct treatment, such as irradiating a tumor. Traditionally, this is done through a time-consuming and error-prone process in which an experienced scientist or physician marks a series of parallel contours that outline the structures of interest. Recent advances in surface reconstruction algorithms have led to methods for reconstructing surfaces from nonparallel contours that could greatly reduce the manual component of this process. Despite these technological advances, the segmentation process has remained unchanged. This dissertation takes the first steps toward bridging the gap between the new surface reconstruction technologies and bringing those methods to use in clinical practice. We develop VolumeViewer, a novel interface for modeling surfaces from volume data by allowing the user to sketch contours on arbitrarily oriented cross-sections of the volume. We design the algorithms necessary to support nonparallel contouring, and we evaluate the system with medical professionals using actual patient data. In this way, we begin to understand how nonparallel contouring can aid the segmentation process and expose the challenges associated with a nonparallel contouring system in practice.

A guided approach to segmentation of volumetric data

classroom use is granted without fee provided that copies are not made or distributed for commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the Owner/Author. SIGGRAPH 2014, August 10 – 14, 2014, Vancouver, British Columbia, Canada. 2014 Copyright held by the Owner/Author. ACM 978-1-4503-2958-3/14/08 A Guided Approach to Segmentation of Volumetric Data

User studies on the feasibility of oblique contouring

MRI and CT scanners have long been used to produce three-dimensional samplings of anatomy elements for use in medical visualization and analysis. Physicians often need to construct surfaces representing the anatomical shape in order to conduct treatment, such as radiating a tumor. Traditionally, this is done by a time-consuming process in which an experienced physician marks a series of parallel contours that outline the object of interest.

An interactive tool for fitting surfaces to volume data

MRI and CT scanners have long been used to produce threeWe begin with a volume dataset. In this example, we have a CT scan of a human prostate. There are 256 slices. The figure to the right shows one input slice, along with rectangles that outline the locations of every 10th slice. We perform linear interpolation between each slice in the input dataset, and now we can render arbitrarily oriented image planes. The figure to the right shows one such plane.