Abstract: A computer implemented method of and system for evaluating translations using a trained neural network. The inputs to the neural network for training and evaluating translations include a source content, translated content, source and target language identifier. Additional inputs for neural network training and evaluating translations can include translation origin before adaptation identifiers, file types, translator identifiers, customer identifiers, and content domain identifiers. The neural network outputs an indication that the translation is either correct or incorrect. Corrections by a reviewer or translator can be used in feedback for further training of the evaluation neural network.

Abstract: An exhaust system directs flow of an exhaust gas from an internal combustion engine. The exhaust system includes an exhaust pipe, a first exhaust treatment device configured to introduce a first substance into the exhaust pipe and a second exhaust treatment device configured to introduce a second substance into the exhaust pipe. A switch is operably connected to the first and second exhaust treatment devices and is positionable in a first position and a second position. An injector is operably connected to the switch and the exhaust pipe. With the switch in the first position, the injector is configured to inject the first substance into the exhaust pipe. With the switch in the second position, the injector is configured to inject the second substance into the exhaust pipe.

Abstract: An exhaust system directs flow of an exhaust gas from an internal combustion engine. The exhaust system includes an exhaust pipe, a first exhaust treatment device configured to introduce a first substance into the exhaust pipe and a second exhaust treatment device configured to introduce a second substance into the exhaust pipe. A switch is operably connected to the first and second exhaust treatment devices and is positionable in a first position and a second position. An injector is operably connected to the switch and the exhaust pipe. With the switch in the first position, the injector is configured to inject the first substance into the exhaust pipe. With the switch in the second position, the injector is configured to inject the second substance into the exhaust pipe.

Abstract: A method for transforming a matrix representation of pixels of an image into a vector representation of the image involves partition of the image into contiguous homogeneous regions by applying a watershed type morphologic algorithm to the matrix representation of pixels and identifying the contours of the contiguous homogeneous regions by applying a vector code contouring algorithm of the chain-code type to every region. By selecting quantization level of the gradient, improved image quality, dimensions and noise are achieved.

Abstract: A method for compression includes applying a discrete cosine transform to said image data to obtain transform coefficients, quantizing the transform coefficients by applying a quantization level scaled through a gain value, adjusting the gain value as a function of desired image parameters by executing a first time said quantization operation applying a first gain value and obtaining first quantized data, estimating statistically a second gain value suitable to obtain the desired image parameters, and executing a second time said quantization operation applying said second gain value. The operation of statistical estimation of the second gain value includes evaluating a threshold value as a function of the desired image parameters and setting to zero a percentage of coefficients of the first quantized data as a function of the threshold value.

Abstract: A raster to vector conversion method of an initial digital image including a pixel matrix, includes generating a digital image divided into polygons by dividing the initial digital image into a plurality of base triangles and defining similarity criteria depending on at least one parameter. The conversion method also includes an iterative operation to process the digital image divided into polygons, selecting pairs of polygons adjacent to each other and to satisfy the similarity criteria and merging together the selected polygons.

Abstract: In the method, the membership image is produced through a multi-stage process that contemplates a first punctual classification stage generating a rough membership image, followed by a low-pass filtering of regionalization of the image through a down-sampling of the image followed by an up-sampling of the image, and in a further step of merging the regions that overcomes the superposition effects induced by the low-pass filtering operation. The so generated membership image contains semantic information of the input image that may be used for effectively implementing an adaptive correction/improvement of the color of the input image. For an adaptive correction/improvement of the color, the encoded data of the generated membership image is fed in parallel to the data of the real image to an adaptive modification block of the color of the image pixels.

Abstract: A method for compression includes applying a discrete cosine transform to said image data to obtain transform coefficients, quantizing the transform coefficients by applying a quantization level scaled through a gain value, adjusting the gain value as a function of desired image parameters by executing a first time said quantization operation applying a first gain value and obtaining first quantized data, estimating statistically a second gain value suitable to obtain the desired image parameters, and executing a second time said quantization operation applying said second gain value. The operation of statistical estimation of the second gain value includes evaluating a threshold value as a function of the desired image parameters and setting to zero a percentage of coefficients of the first quantized data as a function of the threshold value.

Abstract: A method for transforming a matrix representation of pixels of an image into a vector representation of the image involves partition of the image into contiguous homogeneous regions by applying a watershed type morphologic algorithm to the matrix representation of pixels and identifying the contours of the contiguous homogeneous regions by applying a vector code contouring algorithm of the chain-code type to every region. By selecting quantization level of the gradient, improved image quality, dimensions and noise are achieved.

Abstract: A raster to vector conversion method of an initial digital image including a pixel matrix, includes generating a digital image divided into polygons by dividing the initial digital image into a plurality of base triangles and defining similarity criteria depending on at least one parameter. The conversion method also includes an iterative operation to process the digital image divided into polygons, selecting pairs of polygons adjacent to each other and to satisfy the similarity criteria and merging together the selected polygons.

Abstract: In the method, the membership image is produced through a multi-stage process that contemplates a first punctual classification stage generating a rough membership image, followed by a low-pass filtering of regionalization of the image through a down-sampling of the image followed by an up-sampling of the image, and in a further step of merging the regions that overcomes the superposition effects induced by the low-pass filtering operation. The so generated membership image contains semantic information of the input image that may be used for effectively implementing an adaptive correction/improvement of the color of the input image. For an adaptive correction/improvement of the color, the encoded data of the generated membership image is fed in parallel to the data of the real image to an adaptive modification block of the color of the image pixels.