A system and method of automatically learning new keywords in a document image based on context such as when a never before seen keyword exists surrounded by other key-value pairs. A machine learning based approach leverages subword embeddings and two-dimensional geometric contexts in a gradient boosted trees classifier. Keys may be composed of multi-word strings or single-word strings.

The disclosure relates to a method and a device for processing an image. The device includes a selecting unit configured to, by recognizing character blocks in the image using a convolutional network classifier or a fully convolutional network classifier, select in the image a seed character block satisfying a condition that a result of recognizing the seed character block is one of elements of a character set composed of characters , , , , , , 0, 1, 2, 3, 4, 5, 6, 7, 8 and 9; and a determining unit configured to determine an area of a middle address of a Japanese recipient address in the image, starting from the seed character block. At least one of the following effects can be achieved by the device and the method: improving efficiency and accuracy of recognizing the middle address of the Japanese recipient address.

There may be provided a processor-implemented method of causing annotation data to be overlaid on a viewport. The method may include: receiving a signal comprising image data, the image data representing a first document; performing optical character recognition on the image data to identify text in the first document; automatically analyzing the text based on stored classification data to identify a first parameter associated with the first document; comparing the first parameter to a second parameter, the second parameter being obtained from a data store and being associated with a second document; determining annotation data based on the comparison, the annotation data determined based on the first parameter and the second parameter; and providing a signal that includes an instruction to cause the annotation data to be overlaid on a viewport displaying a real-time image of the first document.

Systems and methods for image exploration are provided. One aspect of the systems and methods includes identifying a set of images; reducing the set of images to obtain a representative set of images that is distributed throughout the set of images by removing a neighbor image based on a proximity of the neighbor image to an image of the representative set of images; arranging the representative set of images in a grid structure using a self-sorting map (SSM) algorithm; and displaying a portion of the representative set of images based on the grid structure.

An information processing apparatus (100) includes a collation unit (102) that collates first feature information extracted from a person included in a first image (10) with first feature information indicating a feature of a retrieval target person, an extraction unit (104) that extracts second feature information from the person included in the first image in a case where a collation result in the collation unit (102) indicates a match, and a registration unit (106) that stores, in a second feature information storage unit (110), the second feature information extracted from the person included in the first image.

An information processing apparatus (100) includes a collation unit (102) that collates first feature information extracted from a person included in a first image (10) with first feature information indicating a feature of a retrieval target person, an extraction unit (104) that extracts second feature information from the person included in the first image in a case where a collation result in the collation unit (102) indicates a match, and a registration unit (106) that stores, in a second feature information storage unit (110), the second feature information extracted from the person included in the first image.

A mechanism is provided to implement an action-object interaction detection mechanism for recognizing actions in cluttered video scenes. An object bounding box is computed around an object of interest identified in a corresponding label in an initial frame where the object of interest appears in the frame. The object bounding box is propagated from the initial frame to a subsequent frame. For the initial frame and the subsequent frame: the object bounding boxes of the initial frame and the subsequent frame are refined and cropped based on the associated refined object bounding boxes. The set of cropped frames are processed to determine a probability that an action that is to be verified from the corresponding label is being performed. Responsive to determining the probability is equal to or exceeds a verification threshold, a confirmation is provided that the action-object interaction video performs the action that is to be verified.

Systems and methods for lost asset management using photo-matching are disclosed herein. An example method includes capturing a lost asset image corresponding to a lost asset, and generating, by a feature extractor model, a lost asset descriptor that represents features of the lost asset image. The example method also includes storing the lost asset descriptor and the lost asset image in an asset database that includes known asset descriptors, and performing, by a visual search engine, a nearest neighbor search within the asset database to determine a respective metric distance between the lost asset descriptor and the known asset descriptors. The example method also includes determining, by the visual search engine, a ranked list of known assets corresponding to the lost asset, and displaying, at a user interface, the ranked list of known assets for viewing by a user.

Automated analog gauge reading is provided. The method comprises a computer system receiving input of an image and detecting at least one analog gauge in the image. The computer system corrects the orientation of the analog gauge in the image and detects scene text and tick labels on the analog gauge. The computer system determines a position of a pointer on the analog gauge relative to the scene text and outputs a gauge reading value based on an arithmetic progression of tick labels and angle of the pointer with respect to minimum and maximum values on the analog gauge.

Disclosed is a filtering apparatus with respect to an analysis object image. The filtering apparatus includes an image filtering portion configured to determine whether a stored image present in a client is an analysis object image which has a possibility of including a security text, a controlling portion controls transmission of the analysis object image to an analysis server configured to analyze whether the analysis object image includes the security text depending on a result of determination of the image filtering portion, and an interface portion configured to transmit the analysis object image to the analysis server under the control of the controlling portion.