Abstract: Video information may define spherical video content having a duration. Spherical video content may define visual content viewable from a point of view as a function of progress through the spherical video content. Path information may define a path selection for the spherical video content. Path selection may include movement of a viewing window within the spherical video content. The viewing window may define extents of the visual content viewable from the point of view as the function of progress through the spherical video content. Time lapse parameter information may define at least two of a time portion of the duration, an image sampling rate, and a time lapse speed effect. A time lapse video may be generated based on the video information, the path information, and the time lapse parameter information.

Abstract: Video and corresponding metadata is accessed. Events of interest within the video are identified based on the corresponding metadata, and best scenes are identified based on the identified events of interest. A video summary can be generated including one or more of the identified best scenes. The video summary can be generated using a video summary template with slots corresponding to video clips selected from among sets of candidate video clips. Best scenes can also be identified by receiving an indication of an event of interest within video from a user during the capture of the video. Metadata patterns representing activities identified within video clips can be identified within other videos, which can subsequently be associated with the identified activities.

Abstract: Systems and methods for identifying musical features in audio content are presented. Audio content information may be obtained from a digital audio file, the information providing a duration for playback of the audio content and a representation of sound frequencies associated with various moments throughout the duration of the audio content. Sound frequencies associated with one or more of the moments throughout the duration of the audio content may be identified, and characteristics or patterns of the identified sound frequencies may be recognized as being indicative of one or more musical features (e.g., parts, phrases, hits, bars, onbeats, beats, quavers, semiquavers, etc.). Some implementations of the present technology define display objects for display on a digital display, the display objects provided with visual features in an arrangement that distinguishes one musical feature from another across the duration of the audio content.

Abstract: Vehicles such as unmanned air vehicles that are capable of movement from an open, flight configuration to an enclosed configuration in which all major flight components can be protected by an outer shell are disclosed. In the enclosed configuration, the vehicles can take on standard geometric shapes such as a rectangular prism, sphere, cylinder, or another shape, so as to not be recognizable as an unmanned air vehicle. Embodiments of vehicles can also include interchangeable and/or wireless motor arms, motor arms which are electrically connected to the remainder of the vehicle only when in an open configuration, remote controllers removably attached to the remainder of the vehicle, and clip or other attachment mechanisms for attachment to objects such as backpacks.

Abstract: A set of light field sensors may generate light field output signals conveying light field information within fields of view of the set of light field sensors. The generation of the light field outputs signals may be characterized by a subpixel accuracy. The subpixel accuracy may be enabled by a physical link between the set of light field sensors. The fields of view of the set of light field sensors may overlap over an overlap volume. An object may be located within the overlap volume. The light field information characterizing light field emanating from the object may be combined.

Abstract: Audio content may be captured during capture of spherical video content. An audio event within the audio content may indicate an occurrence of a highlight event based on sound(s) originating from audio source(s) captured within an audio event extent within the spherical video content at an audio event moment. Temporal type of the audio event providing guidance with respect to relative temporality of the highlight event with respect to the audio event and spatial type of the audio event providing guidance with respect to relative spatiality of the highlight event with respect to the audio event may be determined. A highlight event moment of the highlight event may be identified based on the audio event moment and temporal type of the audio event. A highlight event extent of the highlight event may be identified based on the audio event extent and the spatial type of the audio event.

Abstract: Multiple videos having individual time durations may be obtained, including a first video with a first time duration. The videos may include visual information defined by one or more electronic media files. An initial portion of the first time duration where the one or more electronic media are to be transcoded may be determined, including determining whether the first time duration is greater than a predefined threshold and if the first time duration is greater than the predefined threshold, determining the initial portion to be an initial time duration that is less than the first time duration. One or more transcoded media files may be generated during the initial portion. A request for the first video may be received from a client computing platform. In response to receipt of the request, the one or more transcoded media files may be transmitted to the client computing platform for display.

Abstract: Methods and apparatus for encoding and decoding image data based on one or more parameters. In one embodiment, various spatial portions or regions of image data (e.g., a still or moving image) are weighted according to the perceived or measured quality. Processing for these weighted regions can be selectively altered or adjusted so as to optimize one or more operational parameters including for example processing and/or memory requirements, or speed.

Abstract: This disclosure relates to systems and methods that generate highlights for a video. A video may be accessed. Criteria for identifying a moment of interest within the video based on a user interaction with a portion of the video may be obtained. Interaction information indicating the user interaction with the portion of the video may be received. The interaction information for the portion of the video may be compared with the criteria. Responsive to the interaction information for the portion of the video indicating the user interaction with the portion of the video satisfying the criteria, a moment in the video corresponding to the portion of the video may be associated with the moment of interest.

Abstract: A gimbal mount system is configured to a couple to a gimbal coupled to and securing a camera. The gimbal mount system includes a handle, a power source, a user interface, a mounting interface, a communication interface, and a communication bus. The mounting interface is located within an end of the gimbal mount system and includes an opening configured to receive a reciprocal mounting protrusion of the gimbal. A locking mechanism removably couples the gimbal to the gimbal mount system. The communication interface is located within the mounting interface and is configured to couple to a reciprocal communication interface of the gimbal. The communication bus is coupled to the power source, user interface, and communication interface and is configured to provide power from the power source to the gimbal. The communication bus may provide instructions to the gimbal based on user input received via the user interface.

Abstract: A method includes receiving, from an imaging device associated with a robotic apparatus, an image signal; identifying an object of interest that is represented in the image signal, wherein the object of interest is identified by the robotic apparatus using the image signal based on irradiation of the object of interest by an external agent; receiving, by the robotic apparatus from the external agent, a task indication that is representative of a task to be performed by the robotic apparatus. Responsive to receipt of the task indication, the method further includes saving a robotic context including information associating the task to be performed with the object of interest, and causing the robotic apparatus to perform the task with respect to the object of interest.

Abstract: Apparatus and methods for digital video data compression via a scalable, multi-resolution approach. In one embodiment, the video content may be encoded using a multi-resolution and/or multi-quality scalable coding approach that reduces computational and/or energy load on a client device. In one implementation, a low fidelity image is obtained based on a first full resolution image. The low fidelity image may be encoded to obtain a low fidelity bitstream. A second full resolution image may be obtained based on the low fidelity bitstream. A portion of a difference image obtained based on the second full resolution image and the first full resolution may be encoded to obtain a high fidelity bitstream. The low fidelity bitstream and the high fidelity bitstream may be provided to e.g., a receiving device.

Abstract: A method is described to greatly improve the efficiency of and reduce the complexity of image compression when using single-sensor color imagers for video acquisition. The method in addition allows for this new image compression type to be compatible with existing video processing tools, improving the workflow for film and television production.

Abstract: A video may be presented on a touchscreen display. Reception of annotation input may be determined based on user's engagement with the touchscreen display. Annotation input may define an in-frame visual annotation for the video. In-frame visual annotation may be associated with a visual portion of the video and one or more points within a duration of the video such that a subsequent presentation of the video includes the in-frame visual annotation positioned at the visual portion of the video at the one or more points. A graphical user interface may be presented on the touchscreen display. The graphical user interface may include one or more animation fields that provide options for selection by the user. The options may define different properties of a moving visual element added to the video. The options may define visual characteristics, presentation periods, and motions of the moving visual element.

Abstract: Disclosed are a system and a method for determining enabling or disabling electronic image stabilization (EIS) for a video frame. An image sensor of a camera system captures a video stream that comprises a plurality of video frames. An image processor determines availability of a computational resource that may process application of EIS on each video frame. Simultaneously, the image processor receives motion data of the camera system from a gyroscope. Based on the computational resource availability, a motion frequency threshold is determined. Based on the gyroscope motion data, a motion frequency of each video frame is estimated. The estimated motion frequency is compared to the determined motion frequency threshold. If the estimated motion frequency is greater than the determined motion frequency threshold, application of EIS is disabled. If the estimated motion frequency is less than or equal to the determined motion frequency threshold, application of EIS is enabled.

Abstract: A camera system with six faces and a front housing is configured to capture images and audio content from external the camera body. The camera system includes an interior audio assembly protected from external environments by a waterproof membrane. The camera system includes drainage ports on the bottom face and the left face of the camera system to encourage moisture to drain from the system. A first drainage channel couples the internal audio assembly to the first drainage port on left face of the camera system and a second drainage channel couples the drainage port on the left face of the camera system to the drainage port on the bottom face of the camera system. A third drainage channel exists between the front face of the camera system and the front housing, the third drainage channel coupling the first and second drainage channels.

Abstract: A camera is configured with multiple microphones to reduce wind noise in audio data collected by the camera. The camera receives motion data, which may comprise data indicating acceleration of the camera, a plurality of video frames received by the camera, or a background level of noise associated with one or more microphones configured on the camera. The camera determines a motion vector from the motion data. The motion vector is parallel to the direction of motion of the camera. The camera selects a subset of one or more microphones in the direction of the motion vector. By recording audio data using the one or more selected microphones, the camera reduces wind noise in the audio data.

Abstract: Users desiring to generate videos from video clips may want to locate moments of interest within the video clips. A system and method described herein may be configured to provide recommendations of moments of interest within video clips post capture of the video clips. User accounts associated with users of the system may include preference information that defines user preferences with respect to values of attributes of video clips. Moments of interest may be identified within individual video clips when the individual video clips have at least one value of at least one attribute specified by the user preferences. Recommendations of identified moments of interest may be provide to users.

Abstract: Apparatus and methods for providing a rotated spherical viewpoint (RSV). In one or more embodiments, the RSV is implemented using equirectangular projections (ERPs). The RSV methodologies described herein are particularly suitable for panning within virtual reality (VR) or 360° panoramic content as there is minimal, if any, geometric distortions at the edge of a given viewport. Moreover, the RSV methodologies described herein may reduce the bandwidth requirements of, for example, devices operating in a client-server architecture. Additionally, the computation requirements for providing RSV are minimal and can be provided using relatively simple rotation transforms.

Abstract: A capture settings for one or more image capture devices may be determined. The capture setting may define one or more aspects of operation for the image capture device(s). The aspect(s) of operation for the image capture device(s) may include one or more aspects of operation for a processor of the image capture device(s), an image sensor of the image capture device(s), and/or an optical element of the image capture device(s). A machine-readable optical code may be generated based on the capture setting and/or other information. The machine-readable optical code may convey the capture setting for the image capture device(s) such that a first image capture device capturing a first image including the machine-readable optical code may: (1) identify the machine-readable optical code within the first image; (2) determine the capture setting conveyed by the machine-readable optical code; and (3) operate in accordance with the capture setting.

Abstract: A mapping system receives sensor data from an unmanned aerial vehicle. The mapping system further receives images from a camera of the unmanned aerial vehicle. The mapping system determines an altitude of the camera based on the sensor data. The mapping system calculates a footprint of the camera based on the altitude of the camera and a field of view of the camera. The mapping system constructs a localized map based on the images and the footprint of the camera.

Abstract: A camera system includes a lens assembly and image processing electronics internal to the camera housing and thermally coupled to the battery assembly. The battery assembly is sensitive to low ambient temperatures and may become damaged if the temperature of the assembly becomes sufficiently low. The camera system comprises a thermal management system that dissipates heat from electronic components of the camera to increase or maintain the temperature of the battery assembly. The thermal management configures the electronic components in different modes to create a step-wise increase of the resistive heat generated in the camera system and thereby increase the temperature of the components in the camera system.

Abstract: Multiple cameras are arranged in an array at a pitch, roll, and yaw that allow the cameras to have adjacent fields of view such that each camera is pointed inward relative to the array. The read window of an image sensor of each camera in a multi-camera array can be adjusted to minimize the overlap between adjacent fields of view, to maximize the correlation within the overlapping portions of the fields of view, and to correct for manufacturing and assembly tolerances. Images from cameras in a multi-camera array with adjacent fields of view can be manipulated using low-power warping and cropping techniques, and can be taped together to form a final image.

Abstract: Video and corresponding metadata is accessed. Events of interest within the video are identified based on the corresponding metadata, and best scenes are identified based on the identified events of interest. In one example, best scenes are identified based on the motion values associated with frames or portions of a frame of a video. Motion values are determined for each frame and portions of the video including frames with the most motion are identified as best scenes. Best scenes may also be identified based on the motion profile of a video. The motion profile of a video is a measure of global or local motion within frames throughout the video. For example, best scenes are identified from portion of the video including steady global motion. A video summary can be generated including one or more of the identified best scenes.

Abstract: Image captured with an image capture device with a rolling shutter may be deformed due to changes in imaging sensor orientation during image capture. Image deformities may occur due to rolling shutter that exposes rows of pixels to light at slightly different times during image capture. Deformities such as wobble, for example, and/or other deformities may be corrected by constructing an output image. The output image may be constructed by determining corresponding pixels within the input image. The location of the input pixel may be determined by performing one or more fixed point iterations to identify one or more input pixels within the input image. A value of the output pixel within the output image may be determined based on a value of a corresponding pixel within the input image.

Abstract: Conventional wireless interface (WiFi) controllers cannot resolve authentication for trusted client devices without calculation from a host processor. Leaving the host processor on or awaking it from a sleep state each time a non-authenticated trusted client device attempts to connect wastes power. A hostless authenticated wake service allows a host controller to enter a sleep state while the WiFi controller responds to multicast domain name service-service discovery (mDNS-SD) queries from trusted client devices. Once a client device is authenticated, the WiFi controller may respond to a trusted client request to awake the host processor for further command processing and service provision. Not only does this approach reduce power consumption by allowing the host processor to remain in the sleep state, it allows trusted client devices to discover its presence while ensuring security.

Abstract: Apparatus and methods for stitching images, or re-stitching previously stitched images. Specifically, the disclosed systems in one implementation save stitching information and/or original overlap source data during an original stitching process. During subsequent retrieval, rendering, and/or display of the stitched images, the originally stitched image can be flexibly augmented, and/or re-stitched to improve the original stitch quality. Practical applications of the disclosed solutions enable, among other things, a user to create and stitch a wide field of view (FOV) panorama from multiple source images on a device with limited processing capability (such as a mobile phone or other capture device). Moreover, post-processing stitching allows for the user to convert from one image projection to another without fidelity loss (or with an acceptable level of loss).

Abstract: A camera detects devices, such as other cameras, smart devices, and access points, with which the camera may communicate. The camera may alternate between operating as a wireless station and a wireless access point. The camera may connect to and receive credentials from a device for another device to which it is not connected. In one embodiment, the camera is configured to operate as a wireless access point, and is configured to receive credentials from a smart device operating as a wireless station. The camera may then transfer the credentials to additional cameras, each configured to operate as wireless stations. The camera and additional cameras may connect to a smart device directly or indirectly (for instance, through an access point), and the smart device may change the camera mode of the cameras. The initial modes of the cameras may be preserved and restored by the smart device upon disconnection.

Abstract: A content visualization system generates visual content for a visualization device based on visual content of a live event. The content visualization system collects visual content and source perspective data from visual content sources. The visualization device requests visual content from the content visualization system by providing device perspective data to the content visualization system. The content visualization system generates visual content for the visualization device based on the visual content from the visual content sources, the source perspective data, and the device perspective data. The content visualization system can determine visual content that is relevant to the device perspective by identifying source perspectives that overlap with the device perspective. The content visualization system generates visual content for the visualization device based on the identified visual content.

Abstract: Systems and methods of distinguishing between feature depicted in an image are presented herein. Information defining an image may be obtained. The image may include visual content comprising an array of pixels. The array may include pixel rows. An identification of a pixel row in an image may be obtained. Distances of individual pixels and/or groups of pixels from the identified row of pixels may be determined. Parameter values for a set of pixel parameters of individual pixels of the image may be determined. Based on one or more of the distances from the identified row of pixels, parameter values of one or more pixel parameters, and/or other information, individual pixels and/or groups of pixels may be classified as one of a plurality of image features.

Abstract: A first portion of a first video segment may be obtained from a user. A second portion of a second video segment may be obtained from the user. The first portion and the second portion may be aggregated to form an aggregated video segment. A first set of capture settings associated with capture of the first portion may be obtained. A second set of capture settings associated with capture of the second portion may be obtained. Preferences for capture settings of an image capturing device may be determined based upon the first and second set of capture settings. Instructions may be transmitted to the image capturing device. The instructions may include the determined preferences for the capture settings and may be configured to cause the image capturing device to adjust the capture settings to the determined preferences.

Abstract: In a video capture system, a virtual lens is simulated when applying a crop or zoom effect to an input video. An input video frame is received from the input video that has a first field of view and an input lens distortion caused by a lens used to capture the input video frame. A selection of a sub-frame representing a portion of the input video frame is obtained that has a second field of view smaller than the first field of view. The sub-frame is processed to remap the input lens distortion to a desired lens distortion in the sub-frame. The processed sub-frame is the outputted.

Abstract: An image including a visual capture of a scene may be accessed. The image may be processed through a convolutional neural network. The convolutional neural network may generate a set of two-dimensional feature maps based on the image. The set of two-dimensional feature maps may be processed through a contextual long short-term memory unit. The contextual long short-term memory unit may generate a set of two-dimensional outputs based on the set of two-dimensional feature maps. A set of attention-masks for the image may be generated based on the set of two-dimensional outputs and the set of two-dimensional feature maps. The set of attention-masks may define dimensional portions of the image. The scene may be classified based on the two-dimensional outputs.

Abstract: Videos may be automatically generated using a set of video clip. Individual moments of interest may be identified within individual video clips of a set of video clips. A moment of interest may correspond to a point in time within a video clip. The point in time may be associated with one or more values of one or more attributes of the video clip. Individual moments of interest may be associated with individual portions of a video. The video may be generated using the set of video clips based on the associations.

Abstract: Systems and method of automatically detecting user responses from users to media content and correlating the user responses to the media content include a determination of an emotional response by a user—for example a smile—to a presentation of media content. The determined emotional response is correlated to the media content. Subsequent presentation of the same media content includes presentation of an indicator—for example a smiley —that indicates the determined emotional response.

Abstract: An image sensor may generate visual output signals conveying visual information within a field of view of the image sensor. First and second video information defining first and video content may be generated based on the visual output signals. The first video content may be captured using a first capture frame rate and the second video content may be captured using a second capture frame rate. Motion of the image sensor and/or motion of an object within the field of view may be determined. The first capture frame rate may be changed based on the determined motion. Third video information defining third video content may be generated based on the first video information and the second video information. The third video content may include one or more frames of the first video content and one or more frames of the second video content.

Abstract: Systems and method of identifying a horizon depicted in an image are presented herein. Information defining an image may be obtained. The image may include visual content comprising an array of pixels. The array may include pixel rows. Parameter values for a set of pixel parameters of individual pixels of the image may be determined. Individual average parameter values of the individual pixel parameters of the pixels in the individual pixel rows may be determined. Based on the average parameter values a pixel row may be identified as depicting a horizon in the image.

Abstract: A compact convolutional neural network may include a preliminary layer group, one or more intermediate layer groups, a final layer group, and/or other layers/layer groups. The preliminary layer group may include an input layer, a first preliminary normalization layer, a preliminary padding layer, a preliminary convolution layer, a preliminary activation layer, a second preliminary normalization layer, and a preliminary downsampling layer. One or more intermediate layer groups may include an intermediate squeeze layer, a first intermediate normalization layer, an intermediate padding layer, a first intermediate expand layer, a second intermediate expand layer, an intermediate concatenation layer, a second intermediate normalization layer, an intermediate activation layer, and an intermediate combination layer.

Abstract: Disclosed is a configuration to control automatic return of an aerial vehicle. The configuration stores a return location in a storage device of the aerial vehicle. The return location may correspond to a location where the aerial vehicle is to return. One or more sensors of the aerial vehicle are monitored during flight for detection of a predefined condition. When a predetermined condition is met a return path program may be loaded for execution to provide a return flight path for the aerial vehicle to automatically navigate to the return location.

Abstract: Disclosed is a system and method for calibrating a magnetometer of a compass. With a global navigation satellite system receiver, a current position is determined. The determined position is used to determine a magnetic inclination (e.g., by a global magnetic field model such as the World Magnetic Model). The calibration system may perform different calibration sequences based on the magnetic inclination. In a first calibration sequence, performed responsive to a determination that a magnetic inclination (or the absolute value of the magnetic inclination) is less than a threshold, magnetic field data is measured by the magnetometer as it is rotated through horizontal rotation paths. If the magnetic inclination is greater than the threshold, magnetic field data is measured by the magnetometer as it is rotated through vertical rotation paths. The measured magnetic field data may be used to determine calibration values for the magnetometer compass.

Abstract: Systems and methods for controlling an unmanned aerial vehicle recognize and interpret gestures by a user. The gestures are interpreted to adjust the operation of the unmanned aerial vehicle, a sensor carried by the unmanned aerial vehicle, or both.

Abstract: A camera includes an input/out system and one or more input/output ports. The camera configures the pins of the input/output port according to a default pin configuration. The camera detects a peripheral device is connected the input/output ports and receives an identifier from the peripheral device indicating whether the peripheral device is a USB3 device or a non-USB3 device. If the peripheral device is a non-USB3 device, the camera remaps the pins to a first configuration. The camera authenticates with the peripheral device to determine if the peripheral device meets a criteria for an approved device. If the authentication is successful, the camera enables communication with the peripheral device and remaps the pins to a second configuration. If the authentication is unsuccessful, the camera disables communication with the peripheral device and remaps the pins of the input/output port to the default configuration.

Abstract: A camera system processes images based on image luminance data. The camera system includes an image sensor, an image pipeline, an encoder and a memory. The image sensor converts light incident upon the image sensor into raw image data. The image pipeline converts raw image data into color-space image data and calculates luminance levels of the color-space image data. The encoder can determine one or more of quantization levels, determining GOP structure or reference frame spacing for the color-space image data based on the luminance levels. The memory stores the color-space image data and the luminance levels.

Abstract: A camera system processes images based on image activity data. The camera system includes an image sensor, an image pipeline, an encoder and a memory. The image sensor converts light incident upon the image sensor into raw image data. The image pipeline converts raw image data into color-space image data and calculates activity variances of the color-space image data. The encoder can determine one or more of quantization levels, block type (Intra vs Inter), determining transform size and type, and determining GOP structure or reference frame spacing for the color-space image data based on the activity variances. The memory stores the color-space image data and the activity variances.

Abstract: A gimbal sleeve for connecting to a camera gimbal may float between a floor surface and a ceiling surface of an aerial vehicle chassis such that the gimbal sleeve has freedom of motion in yaw, pitch, and roll directions relative to the vehicle chassis. The gimbal sleeve may comprise a pair of connection points to the lower dampers on a floor surface of the vehicle chassis and a pair of connection points to upper dampers on a ceiling surface of the vehicle chassis. The connection points include spring forces that enable the gimbal sleeve to return to an equilibrium position in response to external vibrations and reduce the magnitude of vibrations transferred from the aerial vehicle to the gimbal and camera systems.