Abstract: Examples relate to an optical system and to an apparatus, method and computer program for an optical system. The optical imaging system comprises a camera for providing a camera image of a surface of an object. The optical imaging system further comprises one or more measurement modules for performing one or more measurements at one or more points of the surface of the object. The optical imaging system further comprises a display module. The optical imaging system further comprises a processing module configured to obtain the camera image of the surface of the object from the camera. The processing module is configured to obtain the one or more measurements at the one or more points of the surface of the object from the one or more measurement modules. The processing module is configured to determine a spatial location of the one or more points of the surface relative to the camera image.

Abstract: An eye detecting device is configured to: acquire a color image including a face of a person taken by an image taking device; generate a grayscale image by multiplying each of a red component value, a green component value, and a blue component value of each pixel of the color image by a predetermined ratio according to characteristics of a lens of glasses that the person wears; detect an eye of the person from the grayscale image; and output eye information on the eye of the person.

Abstract: An image pre-processing method and an image processing apparatus for a fundoscopic image are provided. A region of interest (ROI) is obtained from a fundoscopic image to generate a first image. The ROI is focused on an eyeball in the fundoscopic image. A smoothing process is performed on the first image to generate a second image. A value difference between neighboring pixels in the second image is increased to generate a third image.

Abstract: In some embodiments, an apparatus includes a memory and a processor. The processor is configured to extract a set of features from a potentially malicious file and provide the set of features as an input to a normalization layer of a neural network. The processor is configured to implement the normalization layer by calculating a set of parameters associated with the set of features and normalizing the set of features based on the set of parameters to define a set of normalized features. The processor is further configured to provide the set of normalized features and the set of parameters as inputs to an activation layer of the neural network such that the activation layer produces an output based on the set of normalized features and the set of parameters. The output can be used to produce a maliciousness classification of the potentially malicious file.

Abstract: A three-dimensional position and posture recognition device speeds estimation of a position posture and a gripping coordinate posture of a gripping target product. The device includes: a sensor unit configured to measure a distance between an image of an object and the object; and a processing unit configured to calculate an object type included in the image, read model data of each object from the external memory, and create structured model data having a resolution set for each object from the model data, generate measurement point cloud data of a plurality of resolutions from information on a distance between an image of the object and the object, perform a K neighborhood point search using the structured model data and the measurement point cloud data, and perform three-dimensional position recognition processing of the object by rotation and translation estimation regarding a point obtained from the K neighborhood point search.

Abstract: The present disclosure provides for crop yield estimation by identifying, via image processing, a field in which a crop is grown; identifying a plurality of regions within the field; identifying, by processing growth metrics via a model, a plurality of data collection points in the plurality of regions, wherein a given data collection point of the plurality of data collection points within a given region of the plurality of regions is identified by multivariate analysis as representative of growing conditions in the given region; receiving in-field data linked to the data collection points of the plurality; and predicting a yield for the crop in the field based on the in-field data.

Abstract: A method includes: inputting an image of a target to each of a first model for estimating a region of an annular face of the target, a second model for estimating a region in an outer circumference of the annular face and a third model for estimating a region in an inner circumference of the annular face, and outputting a region image, as an estimated region of the annular face, the region image being obtained by synthesizing first, second, and third elements. The first element is a region outputted as a result of the first model for calculating a sum set in the synthesizing, the second element is a region outputted as a result of the second model for calculating a sum set in the synthesizing, and the third element is a region outputted as a result of the third model for calculating a difference set in the synthesizing.

Abstract: A method includes the following steps: pre-processing chest X-ray films to obtain initial X-ray film images that meets format requirements; screening the initial X-ray film images to detect whether they are posteroanterior chest images; inputting the posteroanterior chest images into a binary classification model of the deep convolutional neural network for negative and positive classification; inputting the images presenting positive results into a detection model of the deep convolutional neural network to detect a disease type and label an outline of a lesion area in each image; and displaying the disease type and lesion area corresponding to the image.

Abstract: A processor-implemented method implementing a convolution neural network includes: determining a plurality of differential groups by grouping a plurality of raw windows of an input feature map into the plurality of differential groups; determining differential windows by performing, for each respective differential group of the differential groups, a differential operation between the raw windows of the respective differential group; determining a reference element of an output feature map corresponding to a reference raw window among the raw windows by performing a convolution operation between a kernel and the reference raw window; and determining remaining elements of the output feature map by performing a reference element summation operation based on the reference element and each of a plurality of convolution operation results determined by performing respective convolution operations between the kernel and each of the differential windows.

Abstract: A variation of a method for estimating a quantity of a blood component in a fluid canister includes: within an image of a canister, identifying a reference marker on the canister; selecting an area of the image based on the reference marker; correlating a portion of the selected area with a fluid level within the canister; estimating a volume of fluid within the canister based on the fluid level; extracting a feature from the selected area; correlating the extracted featured with a concentration of a blood component within the canister; and estimating a quantity of the blood component within the canister based on the estimated volume and the concentration of the blood component within the canister.

Abstract: Knowledge distillation method for fracture detection includes obtaining medical images including region-level labeled images, image-level diagnostic positive images, and image-level diagnostic negative images, in chest X-rays; performing a supervised pre-training process on the region-level labeled images and the image-level diagnostic negative images to train a neural network to generate pre-trained weights; and performing a semi-supervised training process on the image-level diagnostic positive images using the pre-trained weights. A teacher model is employed to produce pseudo ground-truths (GTs) on the image-level diagnostic positive images for supervising training of a student model, and the pseudo GTs are processed by an adaptive asymmetric label sharpening (AALS) operator to produce sharpened pseudo GTs to provide positive detection responses on the image-level diagnostic positive images.

Abstract: This disclosure relates to a method and system for extracting information from images of one or more templatized documents. A knowledge graph with a fixed schema based on background knowledge is used to capture spatial and semantic relationships of entities present in scanned document and an adaptive lattice-based approach based on formal concepts analysis (FCA) is used to determine a similarity metric that utilizes both spatial and semantic information to determine if the structure of the scanned document image adheres to any of the known document templates. If a known document template whose structure is closely matching the structure of the scanned document is detected, then an inductive rule learning based approach is used to learn symbolic rules to extract information present in scanned document image and if a new document template is detected, then future scanned document images belonging to new document template are automatically processed using the learnt rules.

Abstract: Approximate modeling of next combined result for stopping text-field recognition in a video stream. In an embodiment, text-recognition results are generated from frames in a video stream and combined into an accumulated text-recognition result. A distance between the accumulated text-recognition result and a next accumulated text-recognition result is estimated based on an approximate model of the next accumulated text-recognition result, and a determination is made of whether or not to stop processing based on this estimated distance. After processing is stopped, the final accumulated text-recognition result may be output.

Abstract: A document processing apparatus including a processor configured to receive data including plural document sets each made up of predetermined plural pages; classify the data into sets; and in a case where the data includes an unclassified page, display a candidate for a set to which the unclassified page belongs.

Abstract: The present disclosure provides a method for determining a malignancy likelihood score for breast tissue of a patient. The method includes receiving a plurality of two-dimensional images of the breast tissue, the two-dimensional images being derived from a three-dimensional image of the breast tissue, for each two-dimensional image, providing the two-dimensional image to a first model including a first trained neural network, and receiving a number of indicators from the first model, each indicator being associated with a two-dimensional image included in the plurality of two-dimensional images, generating a synthetic two-dimensional image based on the number of indicators and at least one of the plurality of two-dimensional images, providing the synthetic two-dimensional image to a second model including a second trained neural network, receiving a malignancy likelihood score from the second model, and outputting a report including the malignancy likelihood score to at least one of a memory or a display.

Abstract: In an example embodiment, machine learning is used to intelligently compress documents to reduce the overall footprint of storing large amounts of files for an organization. Specifically, a document is split into parts, with each part representing a grouping of text or an image. Optical character recognition is performed to identify the text in images. Machine learning techniques are then applied to a part of a document in order to determine how relevant the document is for the organization. The parts that are deemed to be not relevant may then be reduced in size, either by omitting them completely or by summarizing them. This allows for the compression to be tailored specifically to the organization, resulting in the ability to compress or eliminate parts of documents that other organizations might have found relevant (and thus would not have been compressed or eliminated through traditional means).

Abstract: A computerized apparatus for creating a highlight video, including an interface for charting an event by recording information in regard to the event, where the event information is associated with time codes. A camera captures video of the event, where the video is associated with the time codes. The interface also creates triggers including lists of at least one action. An input/output uploads the event information, video, and triggers to a server. A processor produces entries in an action log from the event information, compares the entries in the action log to the triggers to determine if any of the entries in the action log match any of the triggers, collects the time codes of those entries that match the triggers, creates video clips from the video of the event, based on the collected time codes, and compiles the video clips into the highlight video.

Abstract: Systems and methods for detecting textured contact lenses in an iris image are disclosed. Initially, “K” images are taken of an eye in near-infrared light, each for a different positioning of the illuminant. Three-dimensional properties of the hypothetical iris are estimated based on image processing methods. The variability of the estimated three-dimensional properties is calculated, and a classification into two classes is made based on the variability. A low variability denotes that an authentic iris pattern was presented to the sensor, whereas a high variability denotes that the sensor observes an eye wearing a textured contact lens. The systems and methods disclosed allow for detecting presentation attacks to iris recognition systems/sensors based on presentation of an eye wearing a textured contact lens in an automatic and accurate way.

Abstract: A device for identifying a pattern over a first surface includes a photosensitive layer, and a light-guiding layer disposed between the photosensitive layer and the first surface. The photosensitive layer includes a plurality of photosensitive elements, each configured to convert incident light into an electrical signal. The light-guiding layer includes a plurality of light-guiding units, each positionally corresponding to one of the plurality of photosensitive elements. Each light-guiding unit include a plurality of light-transmitting portions, each configured to guide a light beam reflected from a feature of the pattern positionally corresponding thereto to transmit therethrough and reach one of the plurality of photosensitive elements corresponding to the each of the plurality of light-guiding units. The device can be used for identification of fingerprint, as well as other biometric or non-biometric features, and has a reduced thickness and an improved accuracy for pattern identification.

Abstract: Some embodiments include a special-purpose hardware accelerator that can perform specialized machine learning tasks during both training and inference stages. For example, this hardware accelerator uses a systolic array having a number of data processing units (“DPUs”) that are each connected to a small number of other DPUs in a local region. Data from the many nodes of a neural network is pulsed through these DPUs with associated tags that identify where such data was originated or processed, such that each DPU has knowledge of where incoming data originated and thus is able to compute the data as specified by the architecture of the neural network. These tags enable the systolic neural network engine to perform computations during backpropagation, such that the systolic neural network engine is able to support training.