Xerox Corporation

Conduent Business Services, LLC

Naver Corporation

Gabriela Csurka

Boris Chidlovskii

Florent C Perronnin

Luca Marchesotti

Stéphane Clinchant

Jutta Willamowski

Thomas Mensink

Jakob Verbeek

Christopher R Dance

Yves Hoppenot

Marco Bressan

Craig John Saunders

Jose Antonio Rodriguez Serrano

Sandra Skaff

Marc Dymetman

Guillaume M Bouchard

Philippe Weinzaepfel

Claudio Cifarelli

Diane Larlus-Larrondo

Ana Fucs

Herve Poirier

Tommaso Colombino

Julien Ah-Pine

Jerome Pouyadou

Martin Humenberger

Yohann Cabon

Harsimrat Singh Sandhawalia

Julian McAuley

Teofilo E De Campos

Diane Larlus

Fabien Baradel

Rui Hu

Neda Salamati

Yanal Wazaefi

A method for training, using a plurality of training images with corresponding six degrees of freedom camera pose for a given environment and a plurality of reference images, each reference image depicting an object-of-interest in the given environment and having a corresponding two-dimensional to three-dimensional correspondence for the given environment, a neural network to provide visual localization by: for each training image, detecting and segmenting object-of-interest in the training image; generating a set of two-dimensional to two-dimensional matches between the detected and segmented objects-of-interest and corresponding reference images; generating a set of two-dimensional to three-dimensional matches from the generated set of two-dimensional to two-dimensional matches and the two-dimensional to three-dimensional correspondences corresponding to the reference images; and determining localization, for each training image, by solving a perspective-n-point problem using the generated set of two-dimensional to three-dimensional matches.

System and method for training a neural network for visual localization based upon learning objects-of-interest dense match regression

A classification method includes receiving a collection of samples, each sample comprising a multidimensional feature representation. A class label prediction for each sample in the collection is generated with one or more pretrained classifiers. For at least one iteration, each multidimensional feature representation is augmented with a respective class label prediction to form an augmented representation, a set of corrupted samples is generated from the augmented representations, and a transformation that minimizes a reconstruction error for the set of corrupted samples is learned. An adapted class label prediction for at least one of the samples in the collection is generated using the learned transformation and information is output, based on the adapted class label prediction. The method is useful in predicting labels for target samples where there is no access to source domain samples that are used to train the classifier and no access to target domain training data.

Transductive adaptation of classifiers without source data

A domain-adapted classification system and method are disclosed. The method includes mapping an input set of representations to generate an output set of representations, using a learned transformation. The input set of representations includes a set of target samples from a target domain. The input set also includes, for each of a plurality of source domains, a class representation for each of a plurality of classes. The class representations are representative of a respective set of source samples from the respective source domain labeled with a respective class. The output set of representations includes an adapted representation of each of the target samples and an adapted class representation for each of the classes for each of the source domains. A class label is predicted for at least one of the target samples based on the output set of representations and information based on the predicted class label is output.

Adapted domain specific class means classifier

A method and apparatus for classifying an image. In one example, the method may include receiving one or more images associated with a source domain and one or more images associated with a target domain, identifying one or more source domain features based on the one or more images associated with the source domain, identifying one or more target domain features based on the one or more images associated with the target domain, training a conditional maximum mean discrepancy (CMMD) engine based on a difference between the one or more source domain features and the one or more target domain features, applying the CMMD engine to the one or more images associated with the target domain to generate one or more labels for each unlabeled target image of the one or more images associated with the target domain and classifying each one of the one or more images in the target domain using the one or more labels.

Conditional adaptation network for image classification

A machine learning method operates on training instances from a plurality of domains including one or more source domains and a target domain. Each training instance is represented by values for a set of features. Domain adaptation is performed using stacked marginalized denoising autoencoding (mSDA) operating on the training instances to generate a stack of domain adaptation transform layers. Each iteration of the domain adaptation includes corrupting the training instances in accord with feature corruption probabilities that are non-uniform over at least one of the set of features and the domains. A classifier is learned on the training instances transformed using the stack of domain adaptation transform layers. Thereafter, a label prediction is generated for an input instance from the target domain represented by values for the set of features by applying the classifier to the input instance transformed using the stack of domain adaptation transform domains.

Domain adaptation by multi-noising stacked marginalized denoising encoders

In camera-based object labeling, boost classifier &#x192;(x)=&#x3a3;&#x3b2;h(x) is trained to classify an image represented by feature vector x using a target domain training set Dof labeled feature vectors representing images acquired by the same camera and a plurality of source domain training sets D, . . . , Dacquired by other cameras. The training applies an adaptive boosting (AdaBoost) algorithm to generate base classifiers h(x) and weights &#x3b2;. The riteration of the AdaBoost algorithm trains candidate base classifiers h(x) each trained on a training set D&#x222a;D, and selects h(x) from previously trained candidate base classifiers. The target domain training set Dmay be expanded based on a prior estimate of the labels distribution for the target domain. The object labeling system may be a vehicle identification system, a machine vision article inspection system, or so forth.

Domain adaptation for image classification with class priors

Methods, systems and processor-readable media for vehicle classification. In general, one or more vehicles can be scanned utilizing a laser scanner to compile data indicative of an optical profile of the vehicle(s). The optical profile associated with the vehicle(s) is then pre-processed. Particular features are extracted from the optical profile following pre-processing of the optical profile. The vehicle(s) can be then classified based on the particular features extracted from the optical feature. A segmented laser profile is treated as an image and profile features that integrate the signal in one of the two directions of the image and Fisher vectors which aggregate statistics of local 'patches' of the image are computed and utilized as part of the extraction and classification process.

Vehicle classification from laser scanners using fisher and profile signatures

A method for diagnosis assistance exploits similarity between a new medical case and existing medical cases and experts when embedded in a common embedding space. Different types of queries are provided for, including a query-by-cases and a query-by-experts. These may be associated with different cost structures that encourage the requester to use the query-by-cases first and seek expert assistance if this proves unsuccessful. Depending on whether the query-by-cases or query-by-experts is requested, a subset of the existing cases or experts is identified based on the similarity of their representations, in the embedding space, with a representation of the new case in the embedding space. There may then be provision for communicating the new case to a selected one or more of the subset of experts for the expert to attempt to provide a diagnosis.

Social medical network for diagnosis assistance

A method and system for managing print jobs is disclosed. A received print job is compared with pending print jobs and executed print jobs, wherein the pending print jobs and the executed print jobs are stored in one or more print queues associated with one or more printing systems. Thereafter, one or more pending print jobs are suspended if the one or more pending print jobs are found similar to the received print job based on the comparison; or the received print job is suspended if the received print job is found similar to one or more of the executed print jobs, based on the comparison.

Method and system for managing print jobs

A system and method for classifying vehicles from laser scan data by receiving laser scan data corresponding to multiple vehicles from a laser scanner; extracting vehicle shapes corresponding to the multiple vehicles based on the laser scan data; aligning the vehicle shapes; and generating vehicle profiles based on the aligned vehicle shapes. The system and method can further include aligning the vehicle shapes using sequence kernels, such as global alignment kernels, and constraining the sequence kernels based on determined weights.

Methods and systems for vehicle classification from laser scans using global alignment

Disclosed is a method and system of differential processing a print job including one or more original documents to render an obfuscated version of the print job. According to an exemplary method, the differential process replaces letters of an original document with randomly selected characters of substantially the same size and location as the original document and objects such as images/graphics are replaced with blurred versions of substantially the same size and locations as the objects in the original document. The differential process creates an obfuscated version of the print job which is illegible and useful for further processing where privacy of documents included in the print job is required.

Print path obfuscation method and system for document content analytics assessment

Classification of image regions comprises: recursively partitioning an image into a tree of image regions having the image as a tree root and at least one image patch in each leaf image region of the tree, the tree having nodes defined by the image regions and edges defined by pairs of nodes connected by edges of the tree; assigning unary classification potentials to nodes of the tree; assigning pairwise classification potentials to edges of the tree; and labeling the image regions of the tree of image regions based on optimizing an objective function comprising an aggregation of the unary classification potentials and the pairwise classification potentials.

Consistent hierarchical labeling of image and image regions

A classification system and method enable improvements to classification with nearest class mean classifiers by computing a comparison measure between a multidimensional representation of a new sample and a respective multidimensional class representation embedded into a space of lower dimensionality than that of the multidimensional representations. The embedding is performed with a projection that has been learned on labeled samples to optimize classification with respect to multidimensional class representations for classes which may be the same or different from those used subsequently for classification. Each multidimensional class representation is computed as a function of a set of multidimensional representations of labeled samples, each labeled with the respective class. A class is assigned to the new sample based on the computed comparison measures.

Metric learning for nearest class mean classifiers

An image segmentation method includes generating a hierarchy of regions by unsupervised segmentation of an input image. Each region is described with a respective region feature vector representative of the region. Hierarchical structures are identified, each including a parent region and its respective child regions in the hierarchy. Each hierarchical structure is described with a respective hierarchical feature vector that is based on the region feature vectors of the respective parent and child regions. The hierarchical structures are classified according to a set of predefined classes with a hierarchical classifier component that is trained with hierarchical feature vectors of hierarchical structures of training images. The training images have semantic regions labeled according to the set of predefined classes. The input image is segmented into a plurality of semantic regions based on the classification of the hierarchical structures and optionally also on classification of the individual regions.

Image segmentation using hierarchical unsupervised segmentation and hierarchical classifiers

A method for segmenting an image includes extracting unary potentials for pixels of the input image. These can be based for each of a set of possible labels, on information for a first channel in the image, such as in the visible range of the spectrum. Pairwise potentials are extracted for neighboring pairs of pixels of the image. These can be based on information for a second channel in the image, such as in the infrared range of the spectrum. An objective function is optimized over pixels of the input image to identify labels for the pixels. The objective function is based on a combination of ones of the extracted unary and pairwise potentials. The image is then segmented, based on the identified pixel labels. The method and system can provide an improvement in segmentation over methods which use only the visible information.

Growing community of inventors

Crolles, France

Gabriela Csurka