Papers by Florent Perronnin
HANDWRITTEN WORD SPOTTER USING SYNTHESIZED TYPED QUERIES
A wordspotting system and method are disclosed for processing candidate word images extracted fro... more A wordspotting system and method are disclosed for processing candidate word images extracted from handwritten documents. In response to a user inputting a selected query string, such as a word to be searched in one or more of the handwritten documents, the system automatically generates at least one computer-generated image based on the query string in a selected font or fonts. A model is trained on the computer-generated image (s) and is thereafter used in the scoring the candidate handwritten word images.
Bags of visual context-dependent words for generic visual categorization
Category context models (64) and a universal context model (62) are generated including sums of s... more Category context models (64) and a universal context model (62) are generated including sums of soft co-occurrences of pairs of visual words in geometric proximity to each other in training images (50) assigned to each category and assigned to all categories, respectively. Context information (76) about an image to be classified (70) are generated including sums of soft co-occurrences of pairs of visual words in geometric proximity to each other in the image to be classified.
An efficient approach to semantic segmentation
Abstract We consider the problem of semantic segmentation, ie assigning each pixel in an image to... more Abstract We consider the problem of semantic segmentation, ie assigning each pixel in an image to a set of pre-defined semantic object categories. State-of-the-art semantic segmentation algorithms typically consist of three components: a local appearance model, a local consistency model and a global consistency model. These three components are generally integrated into a unified probabilistic framework.
Document processing system
A document processing method includes identifying a document which has been selected for printing... more A document processing method includes identifying a document which has been selected for printing that includes at least one image. For at least the one image of the identified document, assigning an image class to the image based on at least one feature extracted from the image. The document is assigned to a document category based on the assigned image class of the at least one image. A printing protocol is assigned to the document based on the assigned document category.
Generic visual categorization using weak geometry
In the first part of this chapter we make a general presentation of the bag-of-keypatches approac... more In the first part of this chapter we make a general presentation of the bag-of-keypatches approach to generic visual categorization (GVC). Our approach is inspired by the bag-of-words approach to text categorization. This method is able to identify the object content of natural images while generalizing across variations inherent to the object class. To obtain a visual vocabulary insensitive to viewpoint and illumination, rotation or affine invariant orientation histogram descriptors of image patches are vector quantized.
Clustering using non-negative matrix factorization on sparse graphs
Abstract: Object clustering techniques are disclosed. A nonnegative sparse similarity matrix (36)... more Abstract: Object clustering techniques are disclosed. A nonnegative sparse similarity matrix (36) is constructed for a set of objects (26). Nonnegative factorization (40) of the nonnegative sparse similarity matrix is performed. Objects of the set of objects are allocated to clusters based on factor matrices generated by the nonnegative factorization of the nonnegative sparse similarity matrix.
System and method for object class localization and semantic class based image segmentation
An automated image processing system and method are provided for class-based segmentation of a di... more An automated image processing system and method are provided for class-based segmentation of a digital image. The method includes extracting a plurality of patches of an input image. For each patch, at least one feature is extracted. The feature may be a high level feature which is derived from the application of a generative model to a representation of low level feature (s) of the patch.
Discriminative face recognition
A novel probabilistic deformable model of face mapping was recently introduced and successfully a... more A novel probabilistic deformable model of face mapping was recently introduced and successfully applied to automatic person identification. In this paper, we consider the use of discrimination to improve the performance of this system. It is possible to introduce discriminative information at two different levels: 1) in the face representations and 2) in the deformable model used to match face images. We explore both types of discrimination and compare them in terms of performance and computational complexity.
Large-scale document image retrieval and classification with runlength histograms and binary embeddings
Abstract We present a new document image descriptor based on multi-scale runlength histograms. Th... more Abstract We present a new document image descriptor based on multi-scale runlength histograms. This descriptor does not rely on layout analysis and can be computed efficiently. We show how this descriptor can achieve state-of-the-art results on two very different public datasets in classification and retrieval tasks. Moreover, we show how we can compress and binarize these descriptors to make them suitable for large-scale applications.
PREDICTING THE AESTHETIC VALUE OF AN IMAGE
Abstract: A system and method for determining the aesthetic quality of an image are disclosed. Th... more Abstract: A system and method for determining the aesthetic quality of an image are disclosed. The method includes extracting a set of local features from the image, such as gradient and/or color features and generating an image representation which describes the distribution of the local features. A classifier system is used for determining an aesthetic quality of the image based on the computed image representation.
Color transfer between images through color palette adaptation
An image adjustment includes adapting a universal palette to generate (i) an input image palette ... more An image adjustment includes adapting a universal palette to generate (i) an input image palette statistically representative of pixels of an input image and (ii) a reference image palette statistically representative of pixels of a reference image, and adjusting at least some pixels of the input image to generate adjusted pixels that are statistically represented by the reference image palette.
Asymmetric score normalization for handwritten word spotting system
A method begins by receiving an image of a handwritten item. The method performs a word segmentat... more A method begins by receiving an image of a handwritten item. The method performs a word segmentation process on the image to produce a sub-image and extracts a set of feature vectors from the sub-image.
LARGE-SCALE ASYMMETRIC COMPARISON COMPUTATION FOR BINARY EMBEDDINGS
A system and method for comparing a query object and one or more of a set of database objects are... more A system and method for comparing a query object and one or more of a set of database objects are provided. The method includes providing quantized representations of database objects. The database objects have each been transformed with a quantized embedding function which is the composition of a real-valued embedding function and a quantization function. The query object is transformed to a representation of the query object in a real-valued embedding space using the real-valued embedding function.
DYNAMIC FONT REPLACEMENT
Automated font mapping is performed for one or more document fonts of a document to map the one o... more Automated font mapping is performed for one or more document fonts of a document to map the one or more document fonts to at least one replacement font. The font mapping is limited by at least one document-specific font mapping limitation. The document is rendered using the at least one replacement font.
SYSTEM AND METHOD FOR RECOMMENDING EDUCATIONAL RESOURCES
A recommender system and method is provided, including receiving a request to recommend a course ... more A recommender system and method is provided, including receiving a request to recommend a course of action related to a plurality of current students in accordance with a plurality of constraints and accessing a computer database storing student data that corresponds to the plurality of current students. The student data includes attribute data corresponding to respective students of the plurality of current students for describing at least one attribute related to the respective students.
UNSTRUCTURED DOCUMENT CLASSIFICATION
A document classification method comprises:(i) classifying pages of an input document to generate... more A document classification method comprises:(i) classifying pages of an input document to generate page classifications;(ii) aggregating the page classifications to generate an input document representation, the aggregating not being based on ordering of the pages; and (iii) classifying the input document based on the input document representation. A page classifier for use in the page classifying operation (i) is trained based on pages of a set of labeled training documents having document classification labels.
An introduction to Biometrics Audio and Video-Based Person Authentification
Modeling images as sets of weighted features
An apparatus, method, and computer program product are provided for generating an image represent... more An apparatus, method, and computer program product are provided for generating an image representation. The method includes receiving an input digital image, extracting features from the image which are representative of patches of the image, generating weighting factors for the features based on location relevance data for the image, and weighting the extracted features with the weighting factors to form a representation of the image.
COMPACT SIGNATURE FOR UNORDERED VECTOR SETS WITH APPLICATION TO IMAGE RETRIEVAL
TRAINING A CLASSIFIER BY DIMENSION-WISE EMBEDDING OF TRAINING DATA
A classifier training method and apparatus for training, a linear classifier trained by the metho... more A classifier training method and apparatus for training, a linear classifier trained by the method, and its use, are disclosed. In training the linear classifier, signatures for a set of training samples, such as images, in the form of multi-dimension vectors in a first multi-dimensional space, are converted to a second multi-dimension space, of the same or higher dimensionality than the first multi-dimension space, by applying a set of embedding functions, one for each dimension of the vector space.
Uploads
Papers by Florent Perronnin