Entity

Time filter

Source Type

Saint-Sauveur-en-Rue, France

Marcotegui B.,MINES ParisTech | Hernandez J.,MINES ParisTech | Retornaz T.,A2iA SA
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2011

Ultimate Opening (UO) is a powerful operator based on numerical residues. In a multi-scale framework, it analyzes an image under a series of increasing openings. Contrasted objects are detected when they are filtered out by an opening, generating an important residue. Gradual transitions make this operator underestimate the contrast of blurred objects. In this paper we propose a solution to this problem, integrating series of non-null residues. The resulting operator handles correctly blurred boundaries, without modifying the behavior on sharp transitions. © 2011 Springer-Verlag. Source


Bluche T.,A2iA SA | Ney H.,CNRS LIMSI | Kermorvant C.,RWTH Aachen
Proceedings of the International Conference on Document Analysis and Recognition, ICDAR | Year: 2013

In this paper, we show that learning features with convolutional neural networks is better than using hand-crafted features for handwritten word recognition. We consider two kinds of systems: a grapheme based segmentation and a sliding window segmentation. In both cases, the combination of a convolutional neural network with a HMM outperform a state-of-the art HMM system based on explicit feature extraction. The experiments are conducted on the Rimes database. The systems obtained with the two kinds of segmentation are complementary: when they are combined, they outperform the systems in isolation. The system based on grapheme segmentation yields lower recognition rate but is very fast, which is suitable for specific applications such as document classification. © 2013 IEEE. Source


Patent
A2iA S.A. | Date: 2011-01-26

A system, method, and computer program product for processing of objects are disclosed. A processor coupled to a graphical user interface is configured to display an object. The processor receives input from a user concerning the object, wherein input relates to at least a partial location of the object, as a mouse position close to the object, a line approximately covering the vertical or horizontal extent of the object, or a box approximately covering the object. The processor provides input to a keying module, wherein the keying module processes the received input and provides the input to a recognition engine. The recognition engine is in communication with the keying module. Based on the received input, the recognition engine provides an exact information concerning the received input to the keying module, as an exact location, a recognition result, and a confidence score qualifying the reliability of the recognition results. The keying module generates an enhanced information about the object based on the information received from the recognition engine and predetermined information concerning the object.


The systems and methods of the present disclosure use a mobile device equipped with a camera to capture and preprocess images of objects including financial documents, financial cards, and identification cards, and to recognize information in the images of the objects. The methods include detecting quadrangles in images of an object in an image data stream generated by the camera, capturing a first image, transforming the first image, binarizing the transformed image, recognizing information in the binarized image, and determining the validity of the recognized information. The method also includes communicating with a server of a financial institution or other organization to determine the validity of the recognized information. The mobile device may include a camera, a display to display an image data stream and captured images, a memory to store a configuration file including parameters for the preprocessing and recognition functions, captured images, and software, and a communication unit to communicate with a server of the financial institution or other organization.


The systems and methods of the present disclosure use a mobile device equipped with a camera to capture and preprocess images of objects including financial documents, financial cards, and identification cards, and to recognize information in the images of the objects. The methods include detecting quadrangles in images of an object in an image data stream generated by the camera, capturing a first image, transforming the first image, binarizing the transformed image, recognizing information in the binarized image, and determining the validity of the recognized information. The method also includes communicating with a server of a financial institution or other organization to determine the validity of the recognized information. The mobile device may include a camera, a display to display an image data stream and captured images, a memory to store a configuration file including parameters for the preprocessing and recognition functions, captured images, and software, and a communication unit to communicate with a server of the financial institution or other organization.

Discover hidden collaborations