I am a Ph.D. student at Purdue University, and my interests surround the area of computer graphics. I received my B.S. in computer science at Cornell University and my Masters at Purdue University. I am a member of the CGVLab where my primary research focus is on appearance editing. My advisor is Daniel Aliaga.
My dissertation topic is achieving compensation compliancy on arbitrarily shaped and colored surfaces for appearance editing applications. Appearance editing is the ability to impart different appearances or visualizations -- called target appearances -- onto an object with carefully controlled digital projector light. Compensation compliancy refers to the ability to achieve a target appearance's colors. Failure to achieve an appearance's colors may lead to a misleading visualization. Achieving compensation compliancy is fundamentally restricted by the surface's orientation with the digital projectors used and the surface's reflectance properties (e.g., type of surface (diffuse, specular), surface albedo colors). Thus, my work focuses on techniques to remove these limitations and generate better compensation compliant appearances. Currently, we have developed a framework to optimally place the projectors to maximize the projector light radiance available illuminating the surface as well as explored possibilities in modifying the colors of the target appearance to a more compliant but perceptually similar set of colors.
Contact:
office: LWSN 3151
email: ajlaw at cs dot purdue dot edu
resume
My dissertation topic is achieving compensation compliancy on arbitrarily shaped and colored surfaces for appearance editing applications. Appearance editing is the ability to impart different appearances or visualizations -- called target appearances -- onto an object with carefully controlled digital projector light. Compensation compliancy refers to the ability to achieve a target appearance's colors. Failure to achieve an appearance's colors may lead to a misleading visualization. Achieving compensation compliancy is fundamentally restricted by the surface's orientation with the digital projectors used and the surface's reflectance properties (e.g., type of surface (diffuse, specular), surface albedo colors). Thus, my work focuses on techniques to remove these limitations and generate better compensation compliant appearances. Currently, we have developed a framework to optimally place the projectors to maximize the projector light radiance available illuminating the surface as well as explored possibilities in modifying the colors of the target appearance to a more compliant but perceptually similar set of colors.
Contact:
office: LWSN 3151
email: ajlaw at cs dot purdue dot edu
resume
Research
Appearance and Compensation Compliancy
Achieving an appearance edit requires that a physical surface is capable of being visually modified to the colors of the desired target appearance (e.g., the ideal appearance sought by the observer). This means that a sufficient amount of light radiance from the projectors must be present to produce the target color at each surface point. A desired target appearance may not always be possible given an appearance editing setup. This leads to errors in the resulting appearance (e.g., color shifting) which can lead to misleading visualizations. In this project, we explore ways to improve the compliancy of target appearances. We have developed a framework for discovering an optimal placement for the projector(s). We define an optimal projector location as a balance between achieving compensation compliancy while being projector light radiance efficient. These two objectives can be balanced in a unique way for any given appearance editing application (e.g., virtual restoration of historical artifacts may place high importance on minimizing the light radiance exposed to the surface of the fragile artifacts). We are also exploring ways to generate perceptually similar appearances by modifying the target appearance's colors. Even by placing projectors in optimal locations, there may be an insufficient amount of projector light radiance. In this case, the only other viable solution to achieving a compliant appearance is to change the colors of the target appearance to be more compliant. Our approach attempts to change the colors to be compliant and as perceptually similar as possible to the original set of colors.
Resolution Enhancement for Appearance Editing
The quality of an appearance edit is directly related to the resolution achieved by the projectors available in the setup. Various factors affect a projector's resolution and radiant power, including the projector's distance and orientation to the editing object's surface. For example, a single projector will inevitably illuminate an object's surface at grazing angles. Such a situation reduces the possible resolution at that particular surface area. In this work, we use multiple projectors with superimposed projections in order to improve the resolution possible for the appearance edit. Superimposed projections reduce the likelihood that any surface area is illuminated only at grazing angles and also provides the opportunity to improve the resolution possible at head-on surface locations. We analyze the projector pixel interaction with the object surfaces to improve the quality of our appearance edits and introduce a method to enable quick appearance changes without having to recalculate full compensation images.
Restoring the Past: Virtual Restoration of Real-World Objects
We present a system which virtually restores damaged or historically significant cultural heritage with digital projector light. This allows us to alter the appearance of the object without needing to physically change the object. Since the appearance changes by virtual restoration are not physically permanent, multiple histories of an artifact are also able to be imparted onto the object as well as various visualizations to analyze the artifact (e.g., highlighting areas of deterioration, pointing out artistic features). Our work addresses both creating a synthetic restored appearance of the artifact as well as computing the appropriate light patterns to project which minimize the light radiance exposed onto the artifact. Minimizing the light appearance is vital to ensure that the artifact is not further damaged by the virtual restoration process. Our system has been used to restore several priceless artifacts up to about 1000 years old in collaboration with the Eiteljorg Museum and the Indianapolis Museum of Art.
Single Viewpoint Model Completion of Symmetric Objects
Complete 3D models of objects is important for many applications. In digital inspection, complete models allows users to analyze the entirety of an object. We present an approach to create complete 3D models of symmetric objects from a single viewpoint. During the model completion process, symmetry enables us to avoid fragment alignment (e.g., ICP) and searching for geometry to complete the model since symmetry defines spatial relationships between the geometry of the object. Altogether, our approach avoids or minimizes three difficulties in previous methods: 1) we avoid 3D triangulation, 2) we avoid searching for geometry to complete our models, and 3) we minimize the viewpoint planning problem to the selection a single viewpoint. Our method also includes algorithms to mitigate global deformations due to capture error. Using our method, we are capable of reconstructing entire scenes with multiple objects for digital inspection.
Conference and Journal Publications
- Law, A., Aliaga, D., Sajadi, B., Majumder, A., and Pizlo, Z. Perceptually-Based Appearance Modification for Compliant Appearance Editing, fast-tracked to Computer Graphics Forum (CGF).
- Aliaga, D., Yeung, Y., Law, A., Sajadi, B., and Majumder, A. Fast High-Resolution Appearance Editing Using Superimposed Projections, accepted to ACM Transactions on Graphics (TOG) pending changes.
- Law, A. and Aliaga, D. 2011. Single Viewpoint Model Completion of Symmetric Objects for Digital Inspection. Computer Vision and Image Understanding (CVIU), 115, 5, 603-610. [PDF][MOV][SD-link]
- Law, A., Aliaga, D., and Majumder, A. 2010. Projector Placement Planning for High Quality Visualizations on Real-World Colored Objects. Proceedings IEEE Visualization, IEEE Transactions on Visualization and Computer Graphics (TVCG), 16, 6, 1633-1641. [PDF][MOV]
- Law, A., Aliaga, D., Yeung, Y., McCoy, R., McKune, A., and Zimmerman, L. 2009. Projecting Restorations in Real-Time for Real-World Objects. Trant, J. and Bearman, D. (eds). Proceedings Museums and the Web, 8 pages. [PDF][MW-link]
- Aliaga, D., Law, A., and Yeung, Y. 2008. A Virtual Restoration Stage for Real-World Objects. Proceedings ACM SIGGRAPH Asia, ACM Transactions on Graphics (TOG), 27, 5. [PDF][MOV] (also featured in newscast "Restoring Art in an Instant" available in Discoveries and Breakthroughs Inside Science (DBIS) [DBIS])
Presentations
- Projector Placement Planning for High Quality Visualizations on Real-World Colored Objects, IEEE Visualization, Salt Lake City, UT, October 2010.
- Projecting Restorations in Real-Time for Real-World Objects, Museums and the Web, Indianapolis, IN, April 2009.
- Changing the Past: Virtual Restoration of Real-World Objects, demo at Indianapolis Museum of Art, Indianapolis, IN, April 2009 (invited talk).
- Changing the Past: Virtual Restoration of Real-World Objects, demo at Eiteljorg Museum, Indianapolis, IN, April 2009 (invited talk).
- A Virtual Restoration Stage for Real-World Objects, SIGGRAPH Asia, Singapore, December 2008.