Abstract: Techniques are disclosed relating to increasing the amount of training data available to machine learning algorithms. A computer system may access an initial set of training data that specifies a plurality of sequences, each of which may define a set of data values. The computer system may amplify the initial set of training data to create a revised set of training data. The amplifying may include identifying sub-sequences of data values in ones of the plurality of sequences in the initial set of training data and using an inheritance algorithm to create a set of additional sequences of data values, where each one of the set of additional sequences may include sub-sequences of data values from at least two different sequences in the initial set of training data. The computer system may process the set of additional sequences using the machine learning algorithm to train a machine learning model.

Abstract: A video coding device may be configured to perform transform data coding according to one or more of the techniques described herein.

Abstract: Methods and systems for calibrating a transmission electron microscope are disclosed. A fiducial mark on the sample holder is used to identify known reference points so that a current collection area and a through-hole on the sample holder can be located. A plurality of beam current and beam area measurements are taken, and calibration tables are extrapolated from the measurements for a full range of microscope parameters. The calibration tables are then used to determine electron dose of a sample during an experiment at a given configuration.

Abstract: A surgical system for use in a surgical procedure is disclosed. The surgical system includes a surgical visualization system and a control circuit configured to propose a portion of an organ to resect based on visualization data from the surgical visualization system, determine a first value of a non-visualization parameter of the organ prior to resection of the portion, and determine a second value of the non-visualization parameter of the organ after resection of the portion. Resection of the portion is configured to yield an estimated capacity reduction of the organ.

Abstract: A vehicular interior rearview mirror assembly includes a mirror head accommodating an interior mirror reflective element. The mirror reflective element has a mirror transflector that transmits near infrared light incident thereon, transmits visible light incident thereon and reflects visible light incident thereon. A camera is disposed within the mirror head and views through the mirror transflector. The camera includes an imaging sensor having a quantum efficiency (QE) of at least 15% for near infrared light having a wavelength of 940 nm. First, second and third near infrared illumination sources are disposed within the mirror head and operable to emit near infrared light that passes through the mirror transflector. The near infrared illumination sources are at respective locations at the mirror reflective element and, when powered, illuminate respective seat regions for a driver-monitoring function or an occupant-detection function.

Abstract: An image decoding method according to the present invention may comprise the steps of: deriving a merge candidate from a candidate block; generating a first merge candidate list including the merge candidate; specifying any one of a plurality of merge candidates included in the first merge candidate list; deriving affine vectors of a current block on the basis of motion information of the specified merge candidate; deriving a motion vector of a sub-block in the current block on the basis of the affine vectors; and performing motion compensation on the sub-block on the basis of the motion vector.

Abstract: An inspection method includes the following steps: identifying a plurality of patterns within an image; and comparing the plurality of patterns with each other for measurement values thereof. The above-mentioned inspection method uses the pattern within the image as a basis for comparison; therefore, measurement values of the plurality of pixels constructing the pattern can be processed with statistical methods and then compared, and the false rate caused by variation of a few pixels is decreased significantly. An inspection system implementing the above-mentioned method is also disclosed.

Abstract: A disclosed system includes a liftgate housing; a tab of the liftgate housing moveable to an acute angle relative to a first portion of the liftgate housing; a tab aperture disposed in the tab; a first wall and a second wall attached to the liftgate housing, where the first wall, the second wall, and the tab form a barrier to prevent debris from entering into the housing; a camera, where at least a portion of the camera is disposed through the tab aperture, and where movement of the tab relative to the first portion of the liftgate housing adjusts the field of view of the camera.

Abstract: An endoscopic visualization system (10) includes an endoscopic subsystem (26) comprising an endoscope having an optical channel and an illumination channel, an image sensor operatively coupled to receive light from the optical channel and a solid-state light source operatively coupled to transmit light through the illumination channel; at least one electrical cable extending from the endoscopic subsystem; and a monitor subsystem (12) comprising a frame defining an inner compartment, a display screen and at least one port configured to receive the at least one electrical cable connected to the endoscopic subsystem.

Abstract: A computer-implemented method for contextually controlling a surgical device is disclosed. The method includes receiving, by a computer system, perioperative data from the surgical device, the perioperative data associated with a surgical procedure; receiving, by the computer system, images from a scope, the images visualizing the surgical device during the surgical procedure; determining, by the computer system, an attribute of the surgical device from the images; determining, by the computer system, procedural context data based at least on the perioperative data and the attribute of the surgical device; and controlling, by the computer system, the surgical device according to the procedural context data.

Abstract: Disclosed are methods, devices, and computer-readable media for detecting lanes and objects in image frames of a monocular camera. In one embodiment, a method is disclosed comprising receiving a plurality of images; identifying a horizon in the plurality of images by inputting the plurality of images into a deep learning (DL) model (either stored on a local device or via a network call); determining one or more camera parameters based on the horizon; and storing or using the camera parameters to initialize a camera.

Abstract: The invention relates to a TOF camera system comprising several cameras, at least one of the cameras being a TOF camera, wherein the cameras are assembled on a common substrate and are imaging the same scene simultaneously and wherein at least two cameras are driven by different driving parameters.

Abstract: A method for endoscopic imaging including: capturing white light images with a video endoscope under white light illumination; evaluating the captured white light images for a structure having a predefined characteristic, when the presence of the structure having the predefined characteristic is found in a white light image, setting a special light imaging mode in which a light source generates special light illumination using the at least one special light and one or more images of a video stream are captured under the special light illumination and subjected to image processing in the set special light processing mode; identifying a subregion of the at least one white light image that contains the structure with the predefined characteristic and reading out only the subregion of a CMOS image sensor associated with the video endoscope, and processing the image data read out from the subregion as one or more special light images.

Abstract: A rearview mirror system for a vehicle providing a direct reflection of scenes in good light and an enhanced display of scenes in bad lighting includes a housing defining an opening, a display screen arranged at the opening, a camera, a fill light lamp, and a controller. The display screen is light-pervious at only one side. The camera and the fill light lamp are arranged on the housing. An image processing chip and the controller are arranged between the display screen and the housing. When ambient brightness is below a preset brightness value, the camera and the fill light lamp are activated, the camera captures images of side and rear scenes of the vehicle and generates an optical signal, the image processing chip, based on the optical signal, outputs an enhanced image on the display screen for better driving safety. A vehicle including the rearview mirror system is also disclosed.

Abstract: Described is a system for evaluating and correcting perception errors in object detection and recognition. The system receives perception data from an environment proximate a mobile platform. Perception probes are generated from the perception data which describe perception characteristics of object detections in the perception data. For each perception probe, probabilistic distributions for true positive and false positive values are determined, resulting in true positive and false negative perception probes. Statistical characteristics of true positive perception probes and false positive perception probes are then determined. Based on the statistical characteristics, true positive perception probes are clustered. An axiom is generated to determine statistical constraints for perception validity for each perception probe cluster. The axiom is evaluated to classify the perception probes as valid or erroneous. Optimal perception parameters are generated by solving an optimization problem based on the axiom.

Abstract: Decoder retrieval timing information, ROI information and tile identification information are conveyed within a video data stream at a level which allows for an easy access by network entities such as MANEs or decoder. In order to reach such a level, information of such types are conveyed within a video data stream by way of packets interspersed into packets of access units of a video data stream. In accordance with an embodiment, the interspersed packets are of a removable packet type, i.e. the removal of these interspersed packets maintains the decoder's ability to completely recover the video content conveyed via the video data stream.

Abstract: An automated surgical system includes a visualization system, an imaging display system, and a surgical hub. The visualization system includes an illumination source and an imaging device. The surgical hub includes a processor and a memory device configured to store instructions. The instructions direct the surgical hub processor to control the illumination source to illuminate a surgical site and to control the imaging device to receive the imaging data. The surgical processor processes the imaging data to create display data that is transmitted over a cloud-based network to the imaging display system for display.

Abstract: A system and method for lightfield communication can be configured to receive one or more images, determine one or more features in the one or more images, and generate a lightfield image from the one or more images where at least one of the one or more features is in a focal plane of the lightfield image.

Abstract: An image signal decoding method according to the present invention comprises the steps of: decoding information indicating whether a current block is encoded using a multi-mode intra prediction; when it is determined that the current block is encoded in the multi-mode intra prediction, dividing the current block into a plurality of partial blocks; and obtaining an intra prediction mode of each of the plurality of partial blocks.

Abstract: Artificial image generation may include obtaining a source image, identifying quantization information from the source image, wherein identifying the quantization information includes identifying multiresolution quantization interval information from the source image, generating a restoration filtered image by restoration filtering the source image, generating a constrained restoration filtered image by constraining the restoration filtered image based on the quantization information, obtaining an unconstrained artificial image based on the constrained restoration filtered image and a generative artificial neural network obtained using a generative adversarial network, obtaining the artificial image by constraining the unconstrained artificial image based on the quantization information, and outputting the artificial image.