Abstract: A method for text-to-image matching includes generating representations of text images, such as license plate images, by embedding each text image into a first vectorial space with a first embedding function. With a second embedding function, a character string, such as a license plate number to be matched, is embedded into a second vectorial space to generate a character string representation. A compatibility is computed between the character string representation and one or more of the text image representations to identify a matching one. The compatibility is computed with a function that uses a transformation which is learned on a training set of labeled images. The learning uses a loss function that aggregates a text-to-image-loss and an image-to-text loss over the training set. The image-to-text loss penalizes the transformation when it correctly ranks a pair of character string representations, given an image representation corresponding to one of them.

Abstract: A system and method for object instance localization in an image are disclosed. In the method, keypoints are detected in a target image and candidate regions are detected by matching the detected keypoints to keypoints detected in a set of reference images. Similarity measures between global descriptors computed for the located candidate regions and global descriptors for the reference images are computed and labels are assigned to at least some of the candidate regions based on the computed similarity measures. Performing the region detection based on keypoint matching while performing the labeling based on global descriptors improves object instance detection.

Abstract: Disclosed is a method and system to learning mid-level features for text images that leverages character bounding box annotations. According to an exemplary embodiment, the disclosed method and system includes extracting semantic local descriptors by aggregating local statistics of small patches and correlating them with character bounding box annotations.

Abstract: An instance-level retrieval method and system are provided. A representation of a query image is embedded in a multi-dimensional space using a learned projection. The projection is learned using category-labeled training data to optimize a classification rate on the training data. The joint learning of the projection and the classifiers improves the computation of similarity/distance between images by embedding them in a subspace where the similarity computation outputs more accurate results. An input query image can thus be used to retrieve similar instances in a database by computing the comparison measure in the embedding space.

Abstract: A training system, training method, and a system and method of use of a trained classification system are provided. A classifier may be trained with a first “cheap” view but not using a second “costly” view of each of the training samples, which is not available at test time. The two views of samples are each defined in a respective original feature space. An embedding function is learned for embedding at least the first view of the training samples into a common feature space in which the second view can also be embedded or is the same as the second view original feature space. Labeled training samples (first view only) for training the classifier are embedded into the common feature space using the learned embedding function. The trained classifier can be used to predict labels for test samples for which the first view has been embedded in the common feature space with the embedding function.

Abstract: An instance-level retrieval method and system are provided. A representation of a query image is embedded in a multi-dimensional space using a learned projection. The projection is learned using category-labeled training data to optimize a classification rate on the training data. The joint learning of the projection and the classifiers improves the computation of similarity/distance between images by embedding them in a subspace where the similarity computation outputs more accurate results. An input query image can thus be used to retrieve similar instances in a database by computing the comparison measure in the embedding space.

Abstract: A training system, training method, and a system and method of use of a trained classification system are provided. A classifier may be trained with a first “cheap” view but not using a second “costly” view of each of the training samples, which is not available at test time. The two views of samples are each defined in a respective original feature space. An embedding function is learned for embedding at least the first view of the training samples into a common feature space in which the second view can also be embedded or is the same as the second view original feature space. Labeled training samples (first view only) for training the classifier are embedded into the common feature space using the learned embedding function. The trained classifier can be used to predict labels for test samples for which the first view has been embedded in the common feature space with the embedding function.

Abstract: 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. Query-dependent estimated distance values are computed for the query object, based on the transformed query object and stored. A comparison (e.g., distance or similarity) measure between the query object and each of the quantized database object representations is computed based on the stored query-dependent estimated distance values. Data is output based on the comparison computation.

Abstract: 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. Query-dependent estimated distance values are computed for the query object, based on the transformed query object and stored. A comparison (e.g., distance or similarity) measure between the query object and each of the quantized database object representations is computed based on the stored query-dependent estimated distance values. Data is output based on the comparison computation.

Abstract: 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. In some such embodiments, the pages of the set of labeled training documents are not labeled, and the page classifier training comprises: clustering pages of the set of labeled training documents to generate page clusters; and generating the page classifier based on the page clusters.