Dr. Gianfranco Doretto
University of Utah, Salt Lake City, Utah, USA
+1
Professor
'%3e%3cpath%20d='M22.1642%2019.0924C21.7317%2018.068%2021.1041%2017.1375%2020.3163%2016.3527C19.5309%2015.5656%2018.6006%2014.9381%2017.5766%2014.5048C17.5674%2014.5002%2017.5582%2014.4979%2017.5491%2014.4933C18.9774%2013.4616%2019.9059%2011.7811%2019.9059%209.88504C19.9059%206.74408%2017.3611%204.19922%2014.2201%204.19922C11.0792%204.19922%208.53429%206.74408%208.53429%209.88504C8.53429%2011.7811%209.46282%2013.4616%2010.8912%2014.4956C10.882%2014.5002%2010.8728%2014.5025%2010.8636%2014.5071C9.83652%2014.9404%208.91487%2015.5617%208.1239%2016.355C7.33685%2017.1404%206.70933%2018.0707%206.276%2019.0947C5.85031%2020.0972%205.62072%2021.172%205.59967%2022.2609C5.59905%2022.2853%205.60335%2022.3097%205.61229%2022.3325C5.62123%2022.3553%205.63465%2022.376%205.65174%2022.3935C5.66883%2022.4111%205.68926%2022.425%205.71182%2022.4345C5.73437%2022.444%205.7586%2022.4489%205.78308%2022.4489H7.15868C7.25956%2022.4489%207.3398%2022.3686%207.3421%2022.2701C7.38795%2020.5001%208.09868%2018.8425%209.35506%2017.5861C10.655%2016.2862%2012.3814%2015.5709%2014.2201%2015.5709C16.0588%2015.5709%2017.7852%2016.2862%2019.0852%2017.5861C20.3415%2018.8425%2021.0523%2020.5001%2021.0981%2022.2701C21.1004%2022.3709%2021.1807%2022.4489%2021.2815%2022.4489H22.6571C22.6816%2022.4489%2022.7058%2022.444%2022.7284%2022.4345C22.751%2022.425%2022.7714%2022.4111%2022.7885%2022.3935C22.8056%2022.376%2022.819%2022.3553%2022.8279%2022.3325C22.8369%2022.3097%2022.8412%2022.2853%2022.8406%2022.2609C22.8176%2021.165%2022.5907%2020.0989%2022.1642%2019.0924ZM14.2201%2013.8284C13.1678%2013.8284%2012.1773%2013.418%2011.4322%2012.6729C10.6871%2011.9278%2010.2767%2010.9374%2010.2767%209.88504C10.2767%208.83271%2010.6871%207.84227%2011.4322%207.09715C12.1773%206.35204%2013.1678%205.94165%2014.2201%205.94165C15.2724%205.94165%2016.2629%206.35204%2017.008%207.09715C17.7531%207.84227%2018.1635%208.83271%2018.1635%209.88504C18.1635%2010.9374%2017.7531%2011.9278%2017.008%2012.6729C16.2629%2013.418%2015.2724%2013.8284%2014.2201%2013.8284Z'%20fill='%238590A2'/%3e%3c/g%3e%3c/svg%3e)
Co-authors
Following
Followers
Research Focus & Metrics
Explore full metrics on Scilit
Academic Career Length
16Years (2011-Present)
(Calculated once Scilit publication data is synced: current year − year of first publication + 1)
Representative Publications
View All Publications1 October 2017
Conference Paper
SMSaeid Motiian
Saeid Motiian
MPMarco Piccirilli
Marco PiccirilliDADonald A. Adjeroh
Donald A. Adjeroh
Gianfranco DorettoThis work provides a unified framework for addressing the problem of visual supervised domain adaptation and generalization with deep models. The main idea is to exploit the Siamese architecture to learn an embedding subspace that is discriminative, and where mapped visual domains are semantically aligned and yet maximally separated. The supervised setting becomes attractive especially when only few target data samples need to be labeled. In this scenario, alignment and separation of semantic probability distributions is difficult because of the lack of data. We found that by reverting to point-wise surrogates of distribution distances and similarities provides an effective solution. In addition, the approach has a high “speed” of adaptation, which requires an extremely low number of labeled target training samples, even one per category can be effective. The approach is extended to domain generalization. For both applications the experiments show very promising results.
1 February 2003 in International Journal of Computer Vision
Journal Article
Gianfranco DorettoACAlessandro Chiuso
Alessandro Chiuso
YWYing Nian Wu
Ying Nian Wu
SSStefano Soatto
Stefano Soatto
1 January 2007
Conference Paper
XWXiaogang Wang
Xiaogang WangGianfranco DorettoTSThomas Sebastian
Thomas SebastianJRJens Rittscher
Jens RittscherPTPeter Tu
Peter TuIn this work we develop appearance models for computing the similarity between image regions containing deformable objects of a given class in realtime. We introduce the concept of shape and appearance context. The main idea is to model the spatial distribution of the appearance relative to each of the object parts. Estimating the model entails computing occurrence matrices. We introduce a generalization of the integral image and integral histogram frameworks, and prove that it can be used to dramatically speed up occurrence computation. We demonstrate the ability of this framework to recognize an individual walking across a network of cameras. Finally, we show that the proposed approach outperforms several other methods.
1 June 2010
Conference Paper
YYYi Yao
Yi Yao
Gianfranco DorettoTransfer learning allows leveraging the knowledge of source domains, available a priori, to help training a classifier for a target domain, where the available data is scarce. The effectiveness of the transfer is affected by the relationship between source and target. Rather than improving the learning, brute force leveraging of a source poorly related to the target may decrease the classifier performance. One strategy to reduce this negative transfer is to import knowledge from multiple sources to increase the chance of finding one source closely related to the target. This work extends the boosting framework for transferring knowledge from multiple sources. Two new algorithms, MultiSource-TrAdaBoost, and TaskTrAdaBoost, are introduced, analyzed, and applied for object category recognition and specific object detection. The experiments demonstrate their improved performance by greatly reducing the negative transfer as the number of sources increases. TaskTrAdaBoost is a fast algorithm enabling rapid retraining over new targets.
1 June 2020
Conference Paper
SPStanislav Pidhorskyi
Stanislav Pidhorskyi
DADonald A. Adjeroh
Donald A. Adjeroh
Gianfranco DorettoAutoencoder networks are unsupervised approaches aiming at combining generative and representational properties by learning simultaneously an encoder-generator map. Although studied extensively, the issues of whether they have the same generative power of GANs, or learn disentangled representations, have not been fully addressed. We introduce an autoencoder that tackles these issues jointly, which we call Adversarial Latent Autoencoder (ALAE). It is a general architecture that can leverage recent improvements on GAN training procedures. We designed two autoencoders: one based on a MLP encoder, and another based on a StyleGAN generator, which we call StyleALAE. We verify the disentanglement properties of both architectures. We show that StyleALAE can not only generate 1024x1024 face images with comparable quality of StyleGAN, but at the same resolution can also produce face reconstructions and manipulations based on real images. This makes ALAE the first autoencoder able to compare with, and go beyond the capabilities of a generator-only type of architecture.
14 January 2011 in Journal of Ambient Intelligence and Humanized Computing
Journal Article
Gianfranco DorettoTSThomas Sebastian
Thomas SebastianPTPeter Tu
Peter TuJRJens Rittscher
Jens RittscherAbout Me
Gianfranco Doretto is a Professor at the University of Utah, with a joint appointment at the Scientific Computing and Imaging Institute and the Department of Biomedical Informatics. He investigates various areas in computer vision and machine learning, focusing on foundational AI challenges and developing deep learning techniques that build robust AI systems for critical fields like healthcare and biomedical applications. He earned his PhD and MS in Computer Science from UCLA in 2005 and 2002, respectively, and also received a master’s degree in Electronics Engineering from the University of Padua, Italy, in 1998. Before joining the University of Utah, Dr. Doretto was a full professor at West Virginia University in Morgantown, in the Department of Computer Science and Electrical Engineering, where he still holds an adjunct professor position. At WVU, he served as director of the AI+Health Engineering Center and led the Vision and Learning Lab. In 2015, he cofounded IstoVisio, the makers of syGlass, a VR software for 3D imaging data visualization, analysis, and education. Before joining WVU, Dr. Doretto was a lead scientist at GE Global Research in Niskayuna, NY, working in the Visualization and Computer Vision Laboratory of the Imaging Technologies Department.
Social Links
