Abstract: A machine learning system may automatically produce classifier algorithms and configuration parameters by selecting them into a set of predetermined unitary algorithms and associated parametrization values. Multiple digital representations of input object items may be produced by varying the position and orientation of the object to be classified and/or of the sensor to capture a digital representation of the object, and/or by varying a physical environment parameter which changes the digital representation capture of the object by the sensor. A robot arm or a conveyor may vary the object and/or the sensor positions and orientations. The machine learning system may employ genetic programming to facilitate the production of classifiers suitable for the classification of multiple digital representations of input object items.

Abstract: A machine learning system may automatically produce classifier algorithms and configuration parameters by selecting them into a set of predetermined unitary algorithms and associated parametrization values. Multiple digital representations of input object items may be produced by varying the position and orientation of the object to be classified and/or of the sensor to capture a digital representation of the object, and/or by varying a physical environment parameter which changes the digital representation capture of the object by the sensor. A robot arm or a conveyor may vary the object and/or the sensor positions and orientations. The machine learning system may employ generic programming to facilitate the production of classifiers suitable for the classification of multiple digital representations of input object items.

Abstract: A machine learning system may automatically produce classifier algorithms and configuration parameters by selecting them into a set of predetermined unitary algorithms and associated parametrization values. Multiple digital representations of input object items may be produced by varying the position and orientation of the object to be classified and/or of the sensor to capture a digital representation of the object, and/or by varying a physical environment parameter which changes the digital representation capture of the object by the sensor. A robot arm or a conveyor may vary the object and/or the sensor positions and orientations. The machine learning system may employ genetic programming to facilitate the production of classifiers suitable for the classification of multiple digital representations of input object items.

Abstract: The present invention proposes a method for generating a security bi-level image used to form one of the inks of a banknote, said image comprising an original bi-level image and a security pattern, said security pattern being obtained in the spatial domain by the inverse Fourier transform of the combination in the frequency domain between the Fourier transform of an auxiliary image and a two-dimensional sweep, said two-dimensional sweep being a circularly symmetric, two-dimensional pattern created by sweeping a self-similar, one-dimensional function along a 360-degree arc, such as said security pattern being detectable from the maximum value of the cross-correlation of said one-dimensional function with the Fourier transform of one line of said banknote, said method comprising the step of: —determining a distance map of the original bi-level image, —generating a merged image by linearly interpolating at least a part of said distance map with said security pattern, —thresholding the merged image to obtain the

Abstract: The present invention proposes a method for generating a security bi-level image used to form one of the inks of a banknote, said image comprising an original bi-level image and a security pattern, said security pattern being obtained in the spatial domain by the inverse Fourier transform of the combination in the frequency domain between the Fourier transform of an auxiliary image and a two-dimensional sweep, said two-dimensional sweep being a circularly symmetric, two-dimensional pattern created by sweeping a self-similar, one-dimensional function along a 360-degree arc, such as said security pattern being detectable from the maximum value of the cross-correlation of said one-dimensional function with the Fourier transform of one line of said banknote, said method comprising the step of:—determining a distance map of the original bi-level image,—generating a merged image by linearly interpolating at least a part of said distance map with said security pattern,—thresholding the merged image to obtain the sec

Abstract: A method to generate a security pattern to be embedded in an original image has the particularity that the detection of the security pattern can be achieved through a simple, low processing capability device. The method for generating a security image comprising the original image and the security pattern is characterized in that the security pattern is formed by an inverse transform of the combination in the frequency domain of an auxiliary image and a radially symmetric, two-dimensional pattern, said two-dimensional pattern created by sweeping a self-similar, one-dimensional function along a 360-degree arc, and the security image is generated by the modulation of at least one color of at least a part of the original image with the pattern.

Abstract: The present application aims to propose a solution for embedding invisible mark suitable for very large production volumes. This is achieved by a method for applying an invisible mark on a media based on a pattern, and comprising the following steps: determining an area to apply this mark, applying a transparent or semi-transparent substance, on the media, said substance being modulated with the active dots to form the pattern by modifying the quantity of the substance at each active dot.

Abstract: The present invention aims to provide a simple and reliable security element into a document, which can be recognized without special equipment. This aim is achieved with a paper or paper-like document with a surface, the surface being provided at least in a partial area with a background layer, wherein at least part of the surface of the document is provided with a security element comprising a series of protuberant structures spaced at regular interval. The protuberant structure creates a relief than can be easily recognized while rubbing the document. Contrarily to watermarks or other stamps, the fact that the structure is spaced at regular interval allows the user to better detect it over random structures. Regular interval does not mean that the space between each protuberant structure is equal. Regular means that each structure has a similar width and each space between the structures has also a similar width.

Abstract: The present invention aims to propose a method to generate a security pattern to be embedded in an original image having the particularity that the detection of the security pattern can be achieved through a simple, low processing capability device. The present invention proposes a method for generating a security image comprising an original image and a security pattern, characterized in that the security pattern has the form of a grating in which the lines width and/or the line spacing is modulated to embody a predefined data, the security image is generated by the modulation of at least one color of at least a part of the original image with the grating. The present invention belongs to the fields of security printing and copy protection of banknotes. It aims at preventing illegal copy and duplication of banknotes by printing a security image on the banknote. A device capable of detecting this security image with one-dimensional signal processing is described.

Abstract: The present application aims to propose a solution for embedding invisible mark suitable for very large production volumes. This is achieved by a method for applying an invisible mark on a media based on a pattern, and comprising the following steps: determining an area to apply this mark, applying a transparent or semi-transparent substance, on the media, said substance being modulated with the active dots to form the pattern by modifying the quantity of the substance at each active dot.