Abstract: In an optical character recognition (OCR) method for digitizing printed text images using a long-short term memory (LSTM) network, text images are pre-processed using a stroke-aware max-min pooling method before being fed into the network, for both network training and OCR prediction. During training, an average stroke thickness is computed from the training dataset. Stroke-aware max-min pooling is applied to each text line image, where minimum pooling is applied if the stroke thickness of the line is greater than the average stroke thickness, while max pooling is applied if the stroke thickness is less than or equal to the average stroke thickness. The pooled images are used for network training. During prediction, stroke-aware max-min pooling is applied to each input text line image, and the pooled image is fed to the trained LSTM network to perform character recognition.

Abstract: Tray selection in printing may involve checking for the availability of paper trays which may provide increased paper throughput, and by postponing a switch to use such trays to help avoid excessive tray switching that could adversely affect paper throughput. When a higher ranked tray is unavailable, tray selection may involve checking the availability of lower ranked trays before the present tray is empty, and by postponing a switch to such trays to help avoid excessive tray switching. A change in printing condition may provide a convenient and opportune time to switch between trays without adversely affecting paper throughput.

Abstract: A method for processing an electronic document (ED) to infer titles in the ED is provided. The method includes: generating a mark-up version of the ED comprising text-styling attributes, text-layout attributes, and text content information of characters included in the ED; generating statistical information of the text-styling and text-layout attributes; calculating, for each text-styling and text-layout attribute, a relative weight score; calculating, for each paragraph in the ED: a styling criteria score and a layout criteria score based on the statistical information and the relative weight scores; a text content score based on the text content information; and a title confidence score based on the styling criteria score, the layout criteria score, and the text content score; and generating a metadata for the ED that includes the title confidence score for each paragraph for use in inferring the titles in the ED.

Abstract: In an optical character recognition (OCR) method for digitizing printed text images using a long-short term memory (LSTM) network, text images are pre-processed using a stroke-aware max-min pooling method before being fed into the network, for both network training and OCR prediction. During training, an average stroke thickness is computed from the training dataset. Stroke-aware max-min pooling is applied to each text line image, where minimum pooling is applied if the stroke thickness of the line is greater than the average stroke thickness, while max pooling is applied if the stroke thickness is less than or equal to the average stroke thickness. The pooled images are used for network training. During prediction, stroke-aware max-min pooling is applied to each input text line image, and the pooled image is fed to the trained LSTM network to perform character recognition.

Abstract: An artificial neural network system for image classification, formed of multiple independent individual convolutional neural networks (CNNs), each CNN being configured to process an input image patch to calculate a classification for the center pixel of the patch. The multiple CNNs have different receptive field of views for processing image patches of different sizes centered at the same pixel. A final classification for the center pixel is calculated by combining the classification results from the multiple CNNs. An image patch generator is provided to generate the multiple input image patches of different sizes by cropping them from the original input image. The multiple CNNs have similar configurations, and when training the artificial neural network system, one CNN is trained first, and the learned parameters are transferred to another CNN as initial parameters and the other CNN is further trained. The classification includes three classes, namely background, foreground, and edge.

Abstract: A method for file conversion includes obtaining an original file including content, that includes objects. The method further includes rendering each object of the multiple objects into a page description language (PDL) format to obtain rendered objects, and extracting, from the original file, a tag for an object of the multiple objects. The method further includes generating object composition metadata based on a composition rule corresponding to the tag. The object composition metadata includes a description of a composition of the object in the original file, and is added with the rendered objects to a PDL file.

Abstract: An artificial neural network system for image classification, including multiple independent individual convolutional neural networks (CNNs) connected in multiple stages, each CNN configured to process an input image to calculate a pixelwise classification. The output of an earlier stage CNN, which is a class score image having identical height and width as its input image and a depth of N representing the probabilities of each pixel of the input image belonging to each of N classes, is input into the next stage CNN as input image. When training the network system, the first stage CNN is trained using first training images and corresponding label data; then second training images are forward propagated by the trained first stage CNN to generate corresponding class score images, which are used along with label data corresponding to the second training images to train the second stage CNN.

Abstract: A gas concentration-length quantification method, including: acquiring a first image including a gas plume with a camera; identifying and segmenting pixels corresponding to the gas plume in the first image; generating a background image corresponding to the first image using an image inpainting algorithm with the first image and positional information of the segmented pixels corresponding to the gas plume as inputs; calculating a gas concentration-length for each pixel corresponding to the gas plume in the first image, based on the first image and the background image data; and triggering an alert when the gas concentration-length for at least one pixel exceeds a threshold level.

Abstract: A method for processing an electronic document with characters includes adjusting the characters to identify lines and words; generating a cluster encompassing all of the lines and the words; setting the cluster as a target; determining whether the target can be divided; in response to determining that the target can be divided, dividing the target into a first plurality of sub-clusters; identifying blocks of text based on the first sub-clusters; and generating a new electronic document with paragraphs and sections based on the blocks of text.

Abstract: A 3D scanning system includes a base stand, two circular arc shaped support tracks, a mounting assembly for mounting the support tracks to the base stand with one or more degrees of rotational freedom, two sensor holders mounted on the respective support track for holding two depth sensors, and a drive mechanism for driving the sensor holders to move along the respective support tracks. The mounting assembly supports relative rotation of the two support tracks and pitch and roll rotations of the support tracks. To perform a 3D scan, a stationary object is placed in front of the two depth sensors. The sensor holders are moved along the respective support tracks to different positions to obtain depth images of the objects from different angles, from which a 3D surface of the object is constructed. Prior to scanning, the two depth sensors are calibrated relative to each other.

Abstract: An image forming apparatus, a method and a computer readable medium for color verification of an image forming apparatus are disclosed. The image forming apparatus includes a sensing device configured to read a color chart; and a processor configured to: perform a first color verification job; perform a second color verification job after the first color verification job when the first color verification job is interrupted, the second color verification job including one or more color patches from a previous color verification job; acquire measurement data from the previous color verification job; acquire measurement data from the second color verification job; and determine whether to re-use the measurement data of the previous color verification job based on one or more differences between the measurement data of the previous color verification job and the measurement data of the second color verification job.

Abstract: Pathological analysis needs instance-level labeling on a histologic image with high accurate boundaries required. To this end, embodiments of the present invention provide a deep model that employs the DeepLab basis and the multi-layer deconvolution network basis in a unified model. The model is a deeply supervised network that allows to represent multi-scale and multi-level features. It achieved segmentation on the benchmark dataset at a level of accuracy which is significantly beyond all top ranking methods in the 2015 MICCAI Gland Segmentation Challenge. Moreover, the overall performance of the model surpasses the state-of-the-art Deep Multi-channel Neural Networks published most recently, and the model is structurally much simpler, more computational efficient and weight-lighted to learn.

Abstract: An image forming apparatus, a method and a computer readable medium for color verification of an image forming apparatus are disclosed. The image forming apparatus includes a sensing device configured to read a color chart; and a processor configured to: perform a first color verification job; perform a second color verification job after the first color verification job when the first color verification job is interrupted, the second color verification job including one or more color patches from a previous color verification job; acquire measurement data from the previous color verification job; acquire measurement data from the second color verification job; and determine whether to re-use the measurement data of the previous color verification job based on one or more differences between the measurement data of the previous color verification job and the measurement data of the second color verification job.

Abstract: A method involves managing electronic documents (EDs). The method includes receiving a first request to convert an original ED including a visible object from a first format to a second format; extracting, in response to the first request, multiple attributes in the original ED specifying the visible object and required to restore the visible object in the first format; generating, by converting the original ED from the first format to the second format, a converted ED including rendering data of the visible object that is grammatically native to the second format and that is necessary to render the visible object from the second format; and embedding the attributes extracted from the original ED into metadata of the converted ED.

Abstract: A method is provided for processing an electronic document (ED), the method includes: generating a mark-up version of the ED that includes a text attribute that covers a first area of the ED and an image attribute that covers a second area of the ED; segmenting the text attribute into lines of text and the lines of text into words separated by spaces; determining whether the second area overlaps with the first area based on at least one of the lines of texts or the words; and displaying, in response to determining that the second area overlaps with the first area, the ED on a display to notify a user that the image attribute is an overlapped objected on the ED.

Abstract: A method and system of calibrating multispectral images from a camera on an aerial vehicle, the method including: capturing multispectral images of an area at a plurality of intervals with a multispectral imaging camera; simultaneously or at an arbitrary time capturing sunlight radiance data for each of the captured images; correlating the images with the sunlight radiance data; and calibrating the multispectral images based on the sunlight radiance data to normalize the multispectral images to one or more previous images of the area.

Abstract: A method for image processing is provided. The method includes: obtaining an image including a table that includes a plurality of cells that each includes a first border, a second border opposite the first border, a third border, and a fourth border opposite the third border; identifying, on a skeleton graph generated for the table, a first non-corner vertex on the first border of a first cell of the plurality of cells; determining a first plurality of splitlines extending from the first non-corner vertex and intersecting with the second border of the first cell without intersecting with the third or fourth border of the first cell; calculating a first plurality of confidence values for the first plurality of splitlines; and splitting the first cell by generating a first artificial edge, based on the first plurality of confidence values and a splitline of the first plurality of splitlines.

Abstract: A method, system, and recording medium are disclosed for verifying a type of print media within trays of an image forming apparatus. The method includes feeding a sheet of a print media from a tray of the image forming apparatus to an in-line spectrophotometer; measuring characteristics of the print media with the in-line spectrophotometer, the measured characteristics of the print media including at least one of brightness, whiteness, or color; and determining if the measured characteristics of the print media matches a type of print media identified in a tray database for the tray of the image forming apparatus.

Abstract: A method, system, and computer readable program code are disclosed for authenticating a user having a mobile device with plural sensors. The method includes recognizing a situation of the mobile device or the user; determining if the situation of the mobile device or the user is a predetermined situation; when the situation is the predetermined situation: determining a first action as an authentication action; determining a single sensor or a first combination of sensors among the plural sensors; and executing a first authentication for the user by using the single sensor or the first combination of sensors; and when the situation is not the predetermined situation: determining a second action as the authentication action; determining a second combination of sensors among the plural sensors; and executing a second authentication for the user by using the second combination of sensors.

Abstract: An improved method and related apparatus for detecting bacteria viability and drug effects using metabolic monitoring. A fluorescent material which is quenched by oxygen is co-localized with the target bacteria, and fluorescence signal is detected at the co-localized places. In some embodiments, the fluorescent material is a fluorescent nanoparticle mixed with the target bacteria in the sample, and co-localization is enhanced using centrifugation, electrophoresis, microflow path modified with antibodies, magnetic force, etc. In some other embodiments, the fluorescent material is a fluorescent film or 3-D matrix immobilized in the bacterial culture chamber, and bacteria in the sample is gathered into localized regions of the bacteria culture chamber where the fluorescent film or 3-D matrix is present by ways of centrifugation, electrophoresis or microflow path. Plasmonic nanoparticles with a metal core and plasmonic film with a metal film may be used as the fluorescent nanoparticles and fluorescent film.