Abstract: The systems and methods provide an action recognition and analytics tool for use in manufacturing, health care services, shipping, retailing and other similar contexts. Machine learning action recognition can be utilized to determine cycles, processes, actions, sequences, objects and or the like in one or more sensor streams. The sensor streams can include, but are not limited to, one or more video sensor frames, thermal sensor frames, infrared sensor frames, and or three-dimensional depth frames. The analytics tool can provide for analyzing ergonomic data from the one or more sensor streams.

Abstract: The systems and methods provide an action recognition and analytics tool for use in manufacturing, health care services, shipping, retailing and other similar contexts. Machine learning action recognition can be utilized to determine cycles, processes, actions, sequences, objects and or the like in one or more sensor streams. The sensor streams can include, but are not limited to, one or more video sensor frames, thermal sensor frames, infrared sensor frames, and or three-dimensional depth frames. The analytics tool can provide for automatic creation of work charts.

Abstract: The systems and methods provide an action recognition and analytics tool for use in manufacturing, health care services, shipping, retailing and other similar contexts. Machine learning action recognition can be utilized to determine cycles, processes, actions, sequences, objects and or the like in one or more sensor streams. The sensor streams can include, but are not limited to, one or more video sensor frames, thermal sensor frames, infrared sensor frames, and or three-dimensional depth frames. The analytics tool can provide for automatic creation of certificates for each instance of a subject product or service. The certificates can string together snippets of the sensor streams along with indicators of cycles, processes, action, sequences, objects, parameters and the like captured in the sensor streams.

Abstract: Embodiments of the present invention provide a machine and continuous data set including process data, quality data, specific actor data, and ergonomic data (among others) to create more accurate job assignments that maximize efficiency, quality and worker safety. Using the data set, tasks may be assigned to actors based on objective statistical data such as skills, task requirements, ergonomics and time availability. Assigning tasks in this way can provide unique value for manufacturers who currently conduct similar analyses using only minimal observational data.

Abstract: The systems and methods provide an action recognition and analytics tool for use in manufacturing, health care services, shipping, retailing and other similar contexts. Machine learning action recognition can be utilized to determine cycles, processes, actions, sequences, objects and or the like in one or more sensor streams. The sensor streams can include, but are not limited to, one or more video sensor frames, thermal sensor frames, infrared sensor frames, and or three-dimensional depth frames. The analytics tool can provide for process validation, anomaly detection and in-process quality assurance.

Abstract: An event detection method can include encoding a plurality of training video snippets into low dimensional descriptors of the training video snippets in a code space. The low dimensional descriptors of the training video snippets can be decoded into corresponding reconstructed video snippets. One or more parameters of the encoding and decoding can be adjusted based on one or more a loss functions to reduce a reconstruction error between the one or more training video snippets and the corresponding one or more reconstructed video snippets, to reduce a class entropy of the plurality of event classes of the code space, to increase fit of the training video snippet, and/or to increase compactness of the code space. The method can further include encoding one or more labeled video snippets of a plurality of event classes into low dimensional descriptors of the labeled video snippets in the code space.

Abstract: Efficient and effective workspace condition analysis systems and methods are presented. In one embodiment, a method comprises: accessing information associated with an activity space, including information on a newly discovered previously unmodeled entity; analyzing the activity information, including activity information associated with the previously unmodeled entity; forwarding feedback on the results of the analysis, including analysis results for the updated modeled information; and utilizing the feedback in a coordinated path plan check process. In one exemplary implementation the coordinated path plan check process comprises: creating a solid/CAD model including updated modeled information; simulating an activity including the updated modeled information; generating a coordinated path plan for entities in the activity space; and testing the coordinated path plan. The coordinated path plan check process can be a success.

Abstract: Techniques for detecting cycle data can include determining object properties and motion properties in a set of consecutive frames of a sensor stream. The cycle data can be determined from the object properties and motion properties without detecting constituent objects. The object properties and motion properties enable improved detection of cycle data in the presence of different object poses, different positions of the object, partial occlusion of the object, varying illumination, variation in the background, and or the like.

Abstract: Techniques for detecting cycle data can include determining object properties and motion properties in a set of consecutive frames of a sensor stream. The cycle data can be determined from the object properties and motion properties without detecting constituent objects. The object properties and motion properties enable improved detection of cycle data in the presence of different object poses, different positions of the object, partial occlusion of the object, varying illumination, variation in the background, and or the like.

Abstract: The systems and methods provide an action recognition and analytics tool for use in manufacturing, health care services, shipping, retailing and other similar contexts. Machine learning action recognition can be utilized to determine cycles, processes, actions, sequences, objects and or the like in one or more sensor streams. The sensor streams can include, but are not limited to, one or more video sensor frames, thermal sensor frames, infrared sensor frames, and or three-dimensional depth frames. The analytics tool can provide for kitting products, including real time verification of packing or unpacking by action and image recognition.

Abstract: In various embodiments, a method includes receiving one or more sensor streams with an engine. The engine identifies one or more actions that are performed at first and second stations of a plurality of stations within the sensor stream(s). The received sensor stream(s) and identified one or more actions performed at the first and second stations are stored in a data structure. The identified one or more actions are mapped to the sensor stream(s). The engine characterizes each of the identified one or more actions performed at each of the first and second stations to produce determined characterizations thereof. Based on one or more of the determined characterizations, automatically producing a recommendation, either dynamically or post-facto, to move at least one of the identified one or more actions performed at one of the stations to another station to reduce cycle time.

Abstract: The systems and methods provide an action recognition and analytics tool for use in manufacturing, health care services, shipping, retailing, restaurants and other similar contexts. Machine learning action recognition can be utilized to determine cycles, processes, actions, sequences, objects and or the like in one or more sensor streams. The sensor streams can include, but are not limited to, one or more video sensor frames, thermal sensor frames, infrared sensor frames, and or three-dimensional depth frames. The analytics tool can provide for establishing traceability.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, to generate a scannable document. In one aspect, a method includes receiving a scan request, wherein the scan request includes a plurality of text images; for each text image of the plurality of text images: rectifying the text image to generate a text image with parallel image lines, generating a plurality of word bounding boxes that enclose one or more connected components in the text image, wherein each word bounding box is associated with a respective word, and generating, for each respective word in the text image, a plurality of points that represent the respective word; combining the plurality of text images to form a single text document; and providing the combined image as a scannable document.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, to generate a scannable document. In one aspect, a method includes receiving a scan request, wherein the scan request includes a plurality of text images; for each text image of the plurality of text images: rectifying the text image to generate a text image with parallel image lines, generating a plurality of word bounding boxes that enclose one or more connected components in the text image, wherein each word bounding box is associated with a respective word, and generating, for each respective word in the text image, a plurality of points that represent the respective word; combining the plurality of text images to form a single text document; and providing the combined image as a scannable document.

Abstract: Systems and methods are provided for sharing a screen from a mobile device. For example, a method includes receiving an image from a mobile device, performing recognition on the image to identify space-delimited strings, and generating a content graph for the image, the content graph having content nodes that represent at least some of the strings and the content graph having edges that represent a relative position of strings associated with the content nodes connected by the edges. The method may also include repeating the receiving, performing recognition, and generating for a plurality of images, the plurality of images belonging to a session, and generating a combined graph from the plurality of content graphs based on similarity of content nodes between content graphs, the combined graph representing text from the plurality of images in reading order.

Abstract: Implementations relate to relate to rectification of distortion in an image. In some implementations, a method includes extracting edgelets from an image, each edgelet defined by a location of a pixel having an edge depicted in the image and defined by a direction of the edge. The method finds at least one vanishing point in the image based on intersections of pairs of the edges of the edgelets. Each vanishing point is selected based on a plurality of the edgelets in the image having edges approximately aligned with the vanishing point. Found vanishing points are used to transform the image to reduce image distortion.

Abstract: In some implementations of edge-aware smoothing, a method includes determining a boundary map for an input image, where the boundary map associates one of multiple different labels to each pixel of the input image, and the labels indicate one or more edges in the input image. The method determines a set of input pixels of the input image eligible to influence an output pixel of an output image, the output pixel corresponding to a pixel of the input image. A blurred pixel value for the output pixel is determined, where the blurred pixel value is based on the set of input pixels and associated labels from the boundary map that correspond to the set of input pixels. The associated labels are used to reduce blurring in the output image of any of the one or more edges present in the set of input pixels.

Abstract: A computer-implemented method of managing information is disclosed. The method can include receiving a message from a mobile device configured to connect to a mobile device network (the message including a digital image taken by the mobile device and including information corresponding to words), determining the words from the digital image information using optical character recognition, indexing the digital image based on the words, and storing the digital image for later retrieval of the digital image based on one or more received search terms.

Abstract: In general, liveness detection techniques are described for facial recognition. The techniques enable potential detection and mitigation of attempts to authenticate by spoofing. An example method includes determining, by a computing device, a challenge pattern against which to match an authentication input to detect liveness, and displaying, using a display device, a graphical user interface (GUI) including an element and moving the element according to the challenge pattern within the GUI. The method further includes receiving, from an image capture device, at least a first image of a face and a second image of the face, and detecting one or more eye movements based on the first and second images of the face. The method further includes determining whether to deny authentication with respect to accessing one or more functionalities controlled by the computing device.

Abstract: Embodiments for automatic extraction of character ground truth data from images are disclosed. A transcription may be rendered in a plurality of fonts and orientations to obtain a set of candidate word templates with associated character bounding boxes. A word template may be selected from the set of candidate word templates, wherein the selected word template corresponds to a word patch from an image. The character bounding boxes, of the selected word template, may be evaluated in a plurality of orientations about each respective character from the word patch to obtain a set of candidate character templates. For each respective character from the word patch, a character template may be selected from the set of candidate character templates, wherein each selected character template corresponds to the respective character from the word patch.