Skip to main content

Applications

Latest Innovations
Textureshop: Scratching a new surface in photo editing Textureshop has created a new photo editing tool that combines existing techniques for shape-from-shading and texture synthesis to enable a user to texture objects... Read More Textureshop has created a new photo editing tool that combines existing techniques for shape-from-shading and texture synthesis to enable a user to texture objects in a photograph. The Textureshop tool allows user to give an object in a photo the appearance of a different texture. Textureshop added more patch deformations than previous models. Instead of selecting patch-like areas directly from source texture, this program distorts the texture coordinate of each of the pixels in the patch with surface normal, and then samples it from texture, so that the texture fits on the underlying surface. Benefits Easier to use - Existing professional photography editing programs require a long time commitment and appreciable skill to edit a feature of a photo while keeping it on the surface of an object in a scene. Operates on a single image - Methods previously employed to texture synthesis have used a series of images from different viewpoints. This application avoids the need for reconstructing global 3-D mesh, and synthesizes the texture instead on a network of individually parameterized surface patches, designed to operate on a single image.
Antenna Selection for Wideband MIMO System This invention is method of pre-computing the optimal pairing of antennas to RF/ analog frontends in wireless MIMO systems without the need for computationally... Read More This invention is method of pre-computing the optimal pairing of antennas to RF/ analog frontends in wireless MIMO systems without the need for computationally complex and power consuming searches. Other MIMO antenna selection schemes demand an exponential computational order for an exhaustive search, and slightly less for sub-optimal selection. This new technique requires zero real-time computations and minimal memory for near-optimal anenna selection. When mathematically expressing the computation requirements for MIMO systems, the relationship of the number of antennas in the device (N) to number of RF frontends (M) is exponential (NM) for exhaustive searches and semi-linear [(N2M) or (N4M)] for iterative/ sub-optimal searches. An antenna selection algorithm that eliminates significant computational requirements, and requires memory of a linear order, defined by the product of the number of antennas and the number of frontends. Estimating the angle spread concisely encapsulates the sources of distortion and interference as clusters. The algorithm then selects the optimal frontend-antenna pair to compensate for those clusters. Applications Home/ Office networking and Wi-Fi: Wireless LANs and WANs will benefit from increased range with fewer dropped connections and faster connection speeds, while users of related technologies like wireless voice over internet protocol (VoIP) would enjoy greater mobility and signal clarity. Bluetooth and other M2M systems: MIMO antenna arrays will be the backbone of countless machine-to-machine devices that rely on intercommunication in order to function. This algorithm would decrease the size of such devices while ensuring continued reliability. Handheld devices: The market demands connectivity in handheld electronics. Portable gaming devices, for instance, require access to community forums, and pocket PCs are equipped with internet browsers and multimedia software to communicate with desktop computers and wireless routers. GPS systems: A constant signal is imperative for high-accuracy GPS devices. This suite of improved antenna technologies could provide for the incorporation of GPS systems into a range of portable devices, which are currently inhibited by size, power, or portability. Cellular communications: These technologies will boost cell phone signal and range, while simultaneously minimizing power consumption and miniaturizing device size. Large or phased arrays of smart antennas: Antennas can be constructed on a large scale to improve long distance communication in cell phone broadcast towers and government/ military base stations. Benefits Increase number of built-in antennas: Use of angle spread estimation allows wireless devices to receive more antenna elements without a corresponding increase in RF frontends. No real-time computational requirements: Wireless devices using this technology would benefit from added signal without forfeiting battery life or device size. The algorithm pre-determines the optimal frontend-antenna pair based on total angle spread, avoiding the cumbersome computational bottleneck of exhaustive or iterative searches for the strongest signal. The exponential computational complexity, present in current antenna selection technologies, quickly becomes prohibitive for moderate-to-large antenna arrays in small devices. Reduce heat load: This algorithm can increase active antenna elements without increasing heat load by eliminating the need for extra processing without increasing number of RF frontends.
Active Interactive Genetic Algorithms for Reducing User Fatigue and Boosting Evaluation Consistency in Search-Based Domain Interactive genetic algorithms (iGAs) depend on the evaluations or decisions provided by a user. The reliability of those evaluations is key for the successful... Read More Interactive genetic algorithms (iGAs) depend on the evaluations or decisions provided by a user. The reliability of those evaluations is key for the successful application of iGAs. However, even moderately-sized problems my require thousands of evaluations that are sometimes between rather similar solutions which can lead to frustration and inconsistent evaluations. This invention provides a scheme for actively combating user fatigue and inconsistency in iGAs. The efficiency gained from using an interactive procedure by requiring fewer evaluations is lost if the user is unable to provide consistent evaluations. This invention leads to the creation of active iGAs (aiGAs) which include steps of learning from human-computer interaction that actively guide the search-based interaction process. The innovation includes an evaluation-relaxation scheme that actively combats user fatigue and inconsistency by addressing two important issues: Lack of computable fitness Lack of a systematic method for modeling user-decision process Active iGAs optimize a synthetic fitness to obtain educated guesses of the user preferences and thereby reduce the number of user evaluations required. Three important components enable this invention to combat user fatigue and inconsistency. Partial Ordering: Qualitative decisions made by the user about relative solution quality are used to generate partial ordering of solutions Induced Complete Order: Concepts of non-domination and domination count from multi-objective evolutionary algorithms to induce a complete order of the solutions in the population based on their partial ordering Surrogate Function: Induced order is used to assign ranks to the solutions and use them Applications This technology provides the ability to solve difficult problems such as data management, engineering design and optimization, and time and resource allocation in large-scale businesses and organizations, more quickly, reliably, and accurately.
Assertion Generation Suite GoldMine software technology is an automated assertion generation tools to aid with the hardware design verification process. It uses data mining and static... Read More GoldMine software technology is an automated assertion generation tools to aid with the hardware design verification process. It uses data mining and static analysis to generate propositional and temporal assertions in Verilog RTL. The suite also includes: A Figure of Merit Evaluation feature - for evaluating and ranking assertions. Code Coverage Evaluations - for estimating the code covered by an assertion. Word Level Assertion Generation tool to make assertions more usable than those generated at bit level. For more information, visit the goldmine website: http://goldmine.csl.illinois.edu/
Crowdsourcing Construction Activity Analysis from Jobsite Video Streams Researchers from the University of Illinois have developed a novel way of addressing the problem of manual workface assessment by developing an intuitive web-based... Read More Researchers from the University of Illinois have developed a novel way of addressing the problem of manual workface assessment by developing an intuitive web-based platform for crowdsourcing the task through massive marketplaces such as Amazon Mechanical Turk. They found that non-expert annotators on Amazon Mechanical Turk were as reliable as construction experts for providing accurate and complete workface assessment results. This technology has the potential to minimize the time needed for workface assessment and allows onsite professionals to focus their time on more important task of root-cause analysis and performance improvements. More information about the Real-time and Automated Monitoring and Control group can be found here: http://raamac.cee.illinois.edu/
Cascade of Recurrent Neural Networks for Image Super Resolution Professor Thomas Huang and his research group have developed a new method of recovering high-resolution images from their low-resolution observations. They... Read More Professor Thomas Huang and his research group have developed a new method of recovering high-resolution images from their low-resolution observations. They developed a Recurrent Neural Network Model for image super-resolution which fully exploits sparse representations prior while keeping all the benefits of deep networks. They trained a cascade of RNNs, each with a fixed upscaling factor, enabling high resolution images upscaled by large ratios to be obtained with less training efforts and better recovery quality This technique had higher Peak Signal to Noise ratio than the best state of the art methods by up to 1dB.
Patci Dr. Vetle Torvik fhas developed a new software tool for identifying scientific articles cited by patents. Using advanced probabilistic models and machine learning... Read More Dr. Vetle Torvik fhas developed a new software tool for identifying scientific articles cited by patents. Using advanced probabilistic models and machine learning, the software tracks down the links to databases containing a particular citation. It can determine the names of articles and the authors cited by the examiner even if the citation does not adhere to any standard format with improved accuracy over existing methods. For more information please see the poster here for details: https://www.ideals.illinois.edu/handle/2142/54885
Visualizations and Animations from the NCSA A number of scientific animations have been developed by the University of Illinois National Center for Supercomputing Applications (NCSA) Advanced Visualization... Read More A number of scientific animations have been developed by the University of Illinois National Center for Supercomputing Applications (NCSA) Advanced Visualization Laboratory (AVL) under the direction of Professor Kalina Borkiewicz. Examples of available animations include the Milky Way, black holes, tornado super-twisters, solar super-storms, and many more. AVL's animations are cinematic virtual tours through astrophysics, earth sciences, engineering, and other data domains. Their work has been shared through venues such as international digital-dome museum shows, high-definition documentary television programs, and IMAX movies. Their work has also been featured in productions such as The Science of Interstellar documentary, The Tree of Life and Europa Report, and the Dynamic Earth IMAX show. For more information, please visit the AVL website. Available Licenses NCSA AVL animations are available non-exclusively for commercial use under the two license agreements provided in the sidebar. Note that the financial terms, statement of attribution, and copyright notice need to be determined and finalized before the license is ready for execution. Please contact OTM at otm@illinois.edu or Professor Kalina Borkiewicz at kalina@illinois.edu. Commercial Use License Producer: a three-party license that is intended for use when an organization engages a separate producer or distributor. Commercial Use License: a two party license that is between the University and an organization
NAMD: Scalable Molecular Dynamics Software NAMD is parallel molecular dynamics software designed to provide high-performance simulations of large biomolecular systems. NAMD was developed by the ... Read More NAMD is parallel molecular dynamics software designed to provide high-performance simulations of large biomolecular systems. NAMD was developed by the Theoretical and Computational Biophysics Group under the direction of Professor Klaus Schulten in collaboration with other faculty and laboratories at the University of Illinois at Urbana-Champaign. To learn more about NAMD, please visit this website. Available Licenses The latest version of NAMD software is available free of charge for use by individuals, companies and academic institutions for non-commercial and internal business use only. The software is available under a ready-to-sign click-through agreement on the NAMD website. From this site, select the “Download NAMD” link and then the appropriate version of the software. You will be prompted to register (or login with your credentials), in order to review and accept the license, and to then receive access to software.
A New Human Face Detection Algorithm and System Thomas Huang has been recognized worldwide as a pioneer in research involving topics such as computer vision and signal processing, winning numerous honors and awards for... Read More Thomas Huang has been recognized worldwide as a pioneer in research involving topics such as computer vision and signal processing, winning numerous honors and awards for his seminal contributions to these fields. One noteable invention is Huang's facial recognition software. Consisting of three modules - face detection, discriminative manifold learning, and multiple linear regression - the age-estimation software was trained on a database containing photos of 1,600 faces. The software can estimate ages from 1 year to 93 years. The software’s accuracy ranges from about 50 percent when estimating ages to within 5 years, to more than 80 percent when estimating ages to within 10 years. The accuracy has the potential to be improved through additional training on larger databases of faces. In addition to performing tasks such as security control and surveillance monitoring, age-estimation software also could be used for electronic customer relationship management. Through facial and emotion recognition algorithms, Huang and his team have also been able to reconstruct a 3D face according to the input image and then personalize the 3D face model. This allows for the possibility of having synthetic talking faces deliver messages, instead of text. A method for detecting an object in an image includes calculating a log L likelihood of pairs of pixels at select positions in the image that are derived from training images. The calculated log L likelihood of the pairs of pixels is compared with a threshold value. The object is detected when the calculated log L likelihood is greater than the threshold value.

Pages