Abstract: An image processing apparatus generates an image indicating a foreground area of a captured image, based on (i) a threshold value corresponding to both luminance difference between the captured image and a background image and luminance of one of the captured image and the background image, (ii) obtained luminance of one of the captured image and the background image, and (iii) obtained luminance difference between the captured image and the background image.

Abstract: The present teaching relates to method, system, medium, and implementations for speech recognition. An audio signal is received that represents a speech of a user engaged in a dialogue. A visual signal is received that captures the user uttering the speech. A first speech recognition result is obtained by performing audio based speech recognition based on the audio signal. Based on the visual signal, lip movement of the user is detected and a second speech recognition result is obtained by performing lip reading based speech recognition. The first and the second speech recognition results are then integrated to generate an integrated speech recognition result.

Abstract: A modular intelligent transportation system, comprising an environmentally protected enclosure, a system communications bus, a processor module, communicating with said bus, having a image data input and an audio input, the processor module analyzing the image data and/or audio input for data patterns represented therein, having at least one available option slot, a power supply, and a communication link for external communications, in which at least one available option slot can be occupied by a wireless local area network access point, having a communications path between said communications link and said wireless access point, or other modular components.

Abstract: Techniques are disclosed for creating a background model of a scene using both a pixel based approach and a context based approach. The combined approach provides an effective technique for segmenting scene foreground from background in frames of a video stream. Further, this approach can scale to process large numbers of camera feeds simultaneously, e.g., using parallel processing architectures, while still generating an accurate background model. Further, using both a pixel based approach and context based approach ensures that the video analytics system can effectively and efficiently respond to changes in a scene, without overly increasing computational complexity. In addition, techniques are disclosed for updating the background model, from frame-to-frame, by absorbing foreground pixels into the background model via an absorption window, and dynamically updating background/foreground thresholds.

Abstract: A multi-configuration massive model system. The system comprises a processor unit and a comparator configured to run on the processor unit, a memory, and a configuration manager. The comparator compares sets of parts for two or more configurations of a vehicle to form a list comprising a group of common parts and a group of unique parts. The memory is configured to store a massive model dataset of the configurations of the vehicle with a list of the group of common parts and the group of unique parts for the configurations of the vehicle. The configuration manager, configured to run on the processor unit, receives input of a selected configuration and performs an action relating to the vehicle using the massive model dataset for the selected configuration of the vehicle with the list of the group of common parts and the groups of unique parts stored in the memory.

Abstract: A method of processing an image by a device obtaining one or more images including captured images of objects in a target space, generating metadata including information about mapping between the one or more images and a three-dimensional (3D) mesh model used to generate a virtual reality (VR) image of the target space, and transmitting the one or more images and the metadata to a terminal.

Abstract: A method and a device for adjusting a 3D model are disclosed. The method includes: fusing a 2D video and a 3D model of a target monitored area to generate a mesh model (S101); determining whether each set plane in the mesh model is even based on the location of a respective reference plane in the 3D model (S102); for each set plane that is uneven, referred to as a target set plane, selecting an uneven point in the target set plane as a mark point (S103), and generating a virtual plane based on the mark point and the reference plane corresponding to the target set plane (S104); and adding the virtual plane to the 3D model, and for the 2D video, saving the 3D model in which the virtual plane is added (S105). With the embodiments of the present application, operations for the uneven planes improve the effect of the fusion of a video and a 3D model. There is no need to modify a 3D scene model, which reduces the labor cost and time cost of modifying the 3D scene model.

Abstract: A method and an apparatus for generating 3D panoramic video are provided. In the method, plural frames are captured from a panoramic video. Each frame is transformed into a polyhedral mapping projection comprising side planes, a top plane and a bottom plane. Displacements of pixels in the side planes are calculated by using the side planes of each frame, and displacements of pixels in the top plane and the bottom plane are calculated by using the displacements of the side planes. Then, the pixels in the side planes, the top plane and the bottom plane of each frame are shifted according the displacements of the polyhedral mapping projection to generate a shifted polyhedral mapping projection. The shifted polyhedral mapping projection is transformed into a shifted frame with 2D space format. The shifted frames and corresponding frames construct 3D images and the 3D images are encoded into a 3D panoramic video.

Abstract: A system and method to insert visual subtitles in videos is described. The method comprises segmenting an input video signal to extract the speech segments and music segments. Next, a speaker representation is associated for each speech segment corresponding to a speaker visible in the frame. Further, speech segments are analyzed to compute the phones and the duration of each phone. The phones are mapped to a corresponding viseme and a viseme based language model is created with a corresponding score. Most relevant viseme is selected for the speech segments by computing a total viseme score. Further, a speaker representation sequence is created such that phones and emotions in the speech segments are represented as reconstructed lip movements and eyebrow movements. The speaker representation sequence is then integrated with the music segments and super imposed on the input video signal to create subtitles.

Abstract: The present invention relates to a system for a cloud streaming service, a method for a still image-based cloud streaming service and an apparatus therefor. The still image-based cloud streaming service can be provided by capturing the changed area in the changed frame by comparing same with the previous frame, selecting a still-image compression technique by considering any one or more from among the size of the changed area and image characteristics, and transmitting, to a user, the changed area encoded by the still-image compression technique. When providing a still image-based cloud streaming service, by utilizing still-image compression techniques which are appropriate to the conditions, the compression efficiency of the still image and speed of the cloud streaming service can be improved.

Abstract: An image recognition method for an image recognition apparatus includes detecting, setting, acquiring, selecting, and specifying. At least one part of an identification target is selected from an identification target image. An inquiry region is set based on the detected part. A feature amount of the set inquiry region is acquired. At least one instance image corresponding to the identification target image is selected based on the acquired feature amount. A specific region of the identification target from the identification target image is specified based on the selected instance image.

Abstract: This disclosure describes techniques and systems for encoding instructions in audio data that, when output on a speaker of a first device in an environment, cause a second device to output content in the environment. In some instances, the audio data has a frequency that is inaudible to users in the environment. Thus, the first device is able to cause the second device to output the content without users in the environment hearing the instructions. In some instances, the first device also outputs content, and the content output by the second device is played at an offset relative to a position of the content output by the first device.

Abstract: An image recognition method for an image recognition apparatus includes detecting, setting, acquiring, selecting, and specifying. At least one part of an identification target is selected from an identification target image. An inquiry region is set based on the detected part. A feature amount of the set inquiry region is acquired. At least one instance image corresponding to the identification target image is selected based on the acquired feature amount. A specific region of the identification target from the identification target image is specified based on the selected instance image.

Abstract: Techniques are disclosed for creating a background model of a scene using both a pixel based approach and a context based approach. The combined approach provides an effective technique for segmenting scene foreground from background in frames of a video stream. Further, this approach can scale to process large numbers of camera feeds simultaneously, e.g., using parallel processing architectures, while still generating an accurate background model. Further, using both a pixel based approach and context based approach ensures that the video analytics system can effectively and efficiently respond to changes in a scene, without overly increasing computational complexity. In addition, techniques are disclosed for updating the background model, from frame-to-frame, by absorbing foreground pixels into the background model via an absorption window, and dynamically updating background/foreground thresholds.

Abstract: Content substitution and/or picture-in-picture technology is used to provide a sign language window that provides a sign language interpretation of audio content in main program audio/video content. The sign language window can be selectively disabled by a user that does not wish to view the sign language video. Also, in some implementations, the user can move the sign language window to a desired location on the display. The desired location may be one of a plurality of pre-defined display locations. This abstract is not to be considered limiting, since other embodiments may deviate from the features described in this abstract.

Abstract: A method of analyzing image data comprises: obtaining a first image of a first part of an object; obtaining a second image of a second part of the object having overlap with the first part; obtaining a mapping between the first and second images; segmenting the second image to obtain a segmentation; detecting outliers in the first image by identifying extreme intensity values of elements within one or more classes of elements on the basis of the segmentation; replacing elements of the second image that correspond to at least some outliers of the first image, with replacement values, to obtain a corrected second image; and updating the segmentation by performing the segmenting on the corrected second image. The detecting outliers, the replacing, and the updating are performed iteratively until a predetermined convergence criterion is met, which represents a point where there is no significant change in the tissue and lesion segmentations.

Abstract: Coordinated operation of a voice-controlled device and an accessory device in an environment is described. A remote system processes audio data it receives from the voice-controlled device in the environment to identify a first intent associated with a first domain, a second intent associated with a second domain, and a named entity associated with the audio data. The remote system sends, to the voice-controlled device, first information for accessing main content associated with the named entity, and a first instruction corresponding to the first intent. The remote system also sends, to the accessory device, second information for accessing control information or supplemental content associated with the main content, and a second instruction corresponding to the second intent. The first and second instructions, when processed by the devices in the environment, cause coordinated operation of the voice-controlled device and the accessory device.

Abstract: Content substitution and/or picture-in-picture technology is used to provide a sign language window that provides a sign language interpretation of audio content in main program audio/video content. The sign language window can be selectively disabled by a user that does not wish to view the sign language video. Also, in some implementations, the user can move the sign language window to a desired location on the display. The desired location may be one of a plurality of pre-defined display locations. This abstract is not to be considered limiting, since other embodiments may deviate from the features described in this abstract.

Abstract: A method and an apparatus for displaying the surroundings of a vehicle and to a driver assistance system having such an apparatus having at least one sensor for producing sensor data for the surroundings of a vehicle. In this case, the sensor data are conditioned to produce raw image data, if need be using a grid model of the surroundings of the vehicle, if need be the raw image data are processed to produce object information using a grid model of the surroundings of the vehicle, and the object information obtained is used to condition raw image data to produce image object data, and finally the image object data are displayed.

Abstract: This disclosure provides a method and system for automatically recognizing facial expressions at variable resolutions of video. According to one exemplary method, facial expressions are detected, extracted and classified from a video sequence based on an automatic localization of the periocular region associated with a detected and extracted face.

Abstract: A computer implemented method of object extraction from video images, the method comprising steps a computer is programmed to perform, the steps comprising: receiving a plurality of video images, deriving a plurality of background templates from at least one of the received video images, calculating a plurality of differences from an individual one of the received video images, each one of the differences being calculated between the individual video image and a respective and different one of the background templates, and extracting an object of interest from the individual video image, using a rule applied on the calculated differences.

Abstract: As the use of facial biometrics expands in the commercial and government sectors, the need to ensure that human facial examiners use proper procedures to compare facial imagery will grow. Human examiners have examined fingerprint images for many years such that fingerprint examination processes and techniques have reached a point of general acceptance for both commercial and governmental use. The growing deployment and acceptance of facial recognition can be enhanced and solidified if new methods can be used to assist in ensuring and recording that proper examination processes were performed during the human examination of facial imagery.

Abstract: An image coding method and apparatus considering human visual characteristics are provided. The image coding method comprises (a) modeling image quality distribution of an input image in units of scenes such that the quality of an image input in units of scenes is gradually lowered from a region of interest to a background region, (b) determining a quantization parameter of each region constituting one scene according to the result of modeling of image quality distribution, (c) quantizing image data in accordance with the quantization parameter, and (d) coding entropy of the quantized image data.

Abstract: An image coding method and apparatus for coding a current block are provided. A first candidate having a first motion vector that has been used to code a first block is derived. A second candidate having a second motion vector and a first reference picture index value that identifies a first reference picture corresponding to the second motion vector is derived. A third candidate having a third motion vector and a second reference picture index value that identifies a second reference picture corresponding to the third motion vector is derived. The second and third motion vectors are first and second zero vectors. The first and second reference picture index values are different. One candidate from a plurality of candidates including the first, second, and third candidates is selected. An index identifying the selected candidate is coded. The selected candidate includes a motion vector and a reference picture index value.

Abstract: A method for improving image quality of image data includes analyzing, for each of a plurality of voxels of image data, a set of entries of a dictionary, wherein an entry represents a mapping between a lower resolution patch of voxels and a corresponding higher resolution patch of voxel or a local neighborhood around a voxel, deriving, for each of the plurality of voxels, a subspace based on the analysis, wherein the subspace is for one of the mapping or the local neighborhood, and restoring target image data based on the subspaces, wherein the target image data is image data with higher image resolution or reduced image noise.

Abstract: A computer implemented method of object extraction from video images, the method comprising steps a computer is programmed to perform, the steps comprising: receiving a plurality of video images, deriving a plurality of background templates from at least one of the received video images, calculating a plurality of differences from an individual one of the received video images, each one of the differences being calculated between the individual video image and a respective and different one of the background templates, and extracting an object of interest from the individual video image, using a rule applied on the calculated differences.

Abstract: Methods, media and devices for generating an optimized image snapshot from a captured sequence of persons participating in a meeting are provided. In some embodiments, methods media and devices for utilizing a captured image as a representative image of a person as a replacement of a video stream; as a representation of a person in offline archiving systems; or as a representation of a person in a system participant roster.

Abstract: An image processing apparatus includes an image obtaining unit that obtains an image of a face, an edge enhancer that performs edge enhancement on the image and forms an edge-enhanced image, a binarizer that performs binarization on the edge-enhanced image and forms a binary image, and an area identifying unit that identifies an eyelash area in the image on the basis of the binary image.

Abstract: Techniques are disclosed for detecting sudden illumination changes using radiance consistency within a spatial neighborhood. A background/foreground (BG/FG) component of a behavior recognition system may be configured to generate a background image depicting a scene background. Further, the (BG/FG) component may periodically evaluate a current video frame to determine whether a sudden illumination change has occurred. A sudden illumination change occurs when scene lighting changes dramatically from one frame to the next (or over a small number of frames).

Abstract: The present invention relates to a method for matching point regions in images with radial distortion that, given two frames acquired by a camera with radial distortion, estimates the global image distortion and the local transformations undergone by the image regions between frames, with the objective of accurately tracking an aligning these image regions in a sequence of frames, calibrating the radial distortion using only moving image points, or estimating the relative change in focal length in cameras with radial distortion and variable zoom using only moving image points, and that comprises the following steps: extracting local image features; tracking local features; and determining the radial distortion calibration based on a computational efficient procedure that uses the information for all local features being tracked.

Abstract: A method for encoding pictures within a groups of pictures using prediction, where a first reference picture from a group of pictures and a second reference pictures from the subsequent group of pictures are used in predicting pictures in the group of pictures associated with the first reference picture. A plurality of anchor pictures in the group of pictures associated with the first reference picture may be predicted using both the first and second reference pictures to ensure a smooth transition between different groups of pictures within a video frame.

Abstract: A business method for using computer image processing for selectable task-based digital video compression is described. The method is intended to reduce travel of experts and let these experts direct field agents in performing tasks remotely. The tasks to be performed in the remote field can be monitoring, manipulating, and navigating. A field agent performs the manipulation and navigation operations, this agent can be a human being or a robot. The task-based compression algorithms use computer vision techniques to extract the bare minimum amount of information from the remote field scenery to allow the task to be performed. High frame rate photo-realistic reconstruction of the remote scene is not generally necessary.