Abstract: In some embodiments, a method receives a first request for a first time window and a first viewport position for a client that is viewing a 360 degree video. A first message is sent to indicate a computing device will push content in a stream for a first one or more time windows other than the first time window. The method pushes content based on the first viewport position for the first time window to the client. A second request is received to change from the first viewport position to a second viewport position. The method sends a second message to indicate the computing device will push content in the stream for a second one or more time windows without canceling the stream. The method pushes content based on the second viewport position in the 360 degree video for the second one or more time windows to the client.

Abstract: In some embodiments, a method maintains a list that is configured to store motion vectors from one or more blocks that have been previously processed in a frame. When a current block is in a first set of positions in the frame, the method resets the list by storing a motion vector with a horizontal component and a vertical component equal to zero in the list. The motion vector with the horizontal component and the vertical component equal to zero from the list is used to transcode the current block. When the current block is not in the first set of positions, the method uses a motion vector from the list to transcode the current block.

Abstract: Systems and techniques are described herein for decreasing quantization latency. In some aspects, a process includes determining a first integer data type of data at least one layer of a neural network is configured to process, and determining a second integer data type of data received for processing by the neural network. The second integer data type can be different than the first integer data type. The process further includes determining a ratio between a first size of the first integer data type and a second size of the second integer data type, and scaling parameters of the at least one layer of the neural network using a scaling factor corresponding to the ratio. The process further includes quantize the scaled parameters of the neural network, and inputting the received data to the neural network with the quantized and scaled parameters.

Abstract: The present disclosure provides a data compression method for quantitative remote sensing with an unmanned aerial vehicle. The method performs preprocessing on a multispectral image acquired by an unmanned aerial vehicle, successively performs a three-dimensional convolution and a two-dimensional convolution on the multispectral image by an encoder to obtain deep feature information, performs quantizing and entropy encoding on the deep feature information, optimally distributes a loss and a code rate of the image through end-to-end joint training to obtain an optimal compressed image, and reconstructs the optimal compressed image by using a decoder. Image reconstruction quality and a compression ratio are improved by performing a plurality of convolutions on a multispectral pattern; quantizing and entropy encoding are performed on the convoluted deep feature information, to remove redundancy in a feature image, so as to improve the image reconstruction quality and the compression ratio.

Abstract: In some embodiments, a method receives session features for a session associated with a request for a video from a client and predicts network conditions for the session using the session features. A subset of available profiles is selected based on the network conditions. The available profiles are associated with a different playback characteristic. The method provides a profile ladder that includes the subset of available profiles for the playback of the video to the client. The profile ladder restricts the client to using the subset of available profiles to request segments of the video during the session.

Abstract: In some embodiments, a method receives a first request for a first time window and a first position. A first message is sent to indicate a computing device will push content to a client in a stream for a video for a first one or more time windows other than the first time window. The method pushes content for the first time window based on the first position. A second request is received to change from the first position to a second position. The method pushes content for a second time window to the client based on the first position. A second message is sent to indicate the computing device will push content to the client in the stream for a second one or more time windows other than the second time window. The method pushes content for a third time window to the client based on the second position.

Abstract: In some embodiments, a method performs a first calculation to select an anchor point based on a first reference block in a first reference frame and a second reference block in a second reference frame. The first reference frame is selected as the anchor point. A third reference block in the second reference frame is selected based on the first reference frame being selected as the anchor point. The method performs a second calculation to select the anchor point based on the first reference block in the first reference frame and the third reference block in the second reference frame and determines whether to use the first reference block and the third reference block as a prediction for a current block based on the second calculation.

Abstract: In some embodiments, a method maintains a first list that stores motion vectors from one or more blocks that have been previously processed in a frame. When a current block is in the first set of positions in the frame, the method resets and uses a motion vector from a second list to transcode the current block due to the resetting of the first list. When the current block is not in the first set of positions, the method uses a motion vector from the first list to transcode the current block.

Abstract: Disclosed are methods and devices for simultaneously detecting surface shapes and thickness distribution of inner and outer walls of a thin-walled rotating body. According to the invention, a chromatic confocal senor head is driven by a bottom rotary table, a main measuring head rotary table and a linear motion shaft mover to perform copy scanning along a surface of a thin-walled shell-type rotating body, so that the detection of the surface shapes and thickness distribution of the inner and outer walls of the thin-walled shell-type rotating body can be realized at the same time. By the method, mechanical interference caused by arranging a measuring head in the thin-walled shell-type rotating body can be avoided, and coordinate consistency of surface shape measurement data and thickness distribution data of the inner and outer walls in a three-dimensional space can be ensured, thus ensuring the overall measurement precision; and meanwhile.

Abstract: An embodiment of a semiconductor package apparatus may include technology to determine a residual error based on coding unit information, and determine a candidate coding unit and an associated rate distortion cost based on the residual error. An embodiment may additionally or alternatively include technology to partition a first coding unit into two or more smaller coding units based on a partition message, accelerate processing of at least one of the two or more smaller coding units, and estimate motion fora frame based at least partially on results of the accelerated processing. Other embodiments are disclosed and claimed.

Abstract: Disclosed are face clustering method and apparatus, image classification storage method, computer-readable storage medium and electronic device.

Abstract: Disclosed are methods and devices for simultaneously detecting surface shapes and thickness distribution of inner and outer walls of a thin-walled rotating body. According to the invention, a chromatic confocal senor head is driven by a bottom rotary table, a main measuring head rotary table and a linear motion shaft mover to perform copy scanning along a surface of a thin-walled shell-type rotating body, so that the detection of the surface shapes and thickness distribution of the inner and outer walls of the thin-walled shell-type rotating body can be realized at the same time. By the method, mechanical interference caused by arranging a measuring head in the thin-walled shell-type rotating body can be avoided, and coordinate consistency of surface shape measurement data and thickness distribution data of the inner and outer walls in a three-dimensional space can be ensured, thus ensuring the overall measurement precision; and meanwhile.

Abstract: The present disclosure discloses a fiber structure, including a plurality of fiber layers. The plurality of the fiber layers are stacked and arranged in a surrounding manner and are impregnated and cured in sequence. The fiber structure of the present disclosure is high in structural strength and low in mass. The present disclosure further provides a gymnastic apparatus force bearing frame, including the above-mentioned fiber structure, which is beneficial to reducing the weight of the gymnastic apparatus force bearing frame, facilitating the movement, transportation, disassembling, and assembling of the apparatus, and reducing the labor burden of an operator.

Abstract: A method includes transcoding a first block from a spatial region. The first block is associated with a first block tree pattern defining a structure of splitting a block into smaller blocks. A bit string of bits for the first block tree pattern is included in an encoded bitstream. The method determines a location of the first block in the spatial region when the first block tree pattern of the first block can be reused for a second block tree pattern of a second block. The spatial region includes blocks and the location is based on the first block being in the spatial region. Information for the second block is included in the encoded bitstream that indicates the location of the first block in the spatial region. The location allows the bit string for the first block tree pattern to be retrieved for use to decode the second block.

Abstract: In some embodiments, after receiving a first set of video segments of a video, a method starts a first computing instance and receives a video segment of the video using the first computing instance. A measurement is calculated based on receiving the video segment. The method uses the measurement to determine whether the first computing instance should continue to be used to receive video segments in the video or should be terminated to not receive video segments in the video.

Abstract: In some embodiments, a method receives a current encoding stream of a video. A first measurement for a reference encoding stream is selected to measure a quality of the reference encoding stream. The method compares a second measurement of the current encoding stream to the first measurement of the reference encoding stream and outputs whether the current encoding stream is validated based on the comparing.

Abstract: Techniques related to motion estimation with neighbor block pattern for video coding.

Abstract: In some embodiments, during playback of a video, using a parameter value of an adaptive bitrate algorithm to analyze playback of the video. The adaptive bitrate algorithm uses the parameter value to select a profile from a plurality of profiles to use to request segments of the video. A method selects a scenario in a plurality of scenarios when a playback condition meets a characteristic value of the scenario. The plurality of scenarios are associated with a different characteristic value. The method adjusts the parameter value of the adaptive bitrate algorithm based on the scenario to generate an adjusted parameter value. The plurality of scenarios have at least one different parameter value. The adjusted parameter value of the adaptive bitrate algorithm is used to analyze the playback of the video.

Abstract: A method includes transcoding a first block of a video. The first block is associated with a first block tree pattern defining a structure of splitting a block into smaller blocks. A bit string of bits for the first block tree pattern is included in an encoded bitstream for the video. The method determines that the first block tree pattern of the first block can be reused for a second block tree pattern of a second block and includes information in the encoded bitstream that indicates that the first block tree pattern is to be used to decode the second block from the encoded bitstream.

Abstract: The present invention includes a reference parallel line creation step of creating a pair of parallel lines that are respectively in contact with outer edges of a workpiece cross-section without traversing the workpiece cross-section and that are located on an extension plane of the workpiece cross-section, a first determination step of comparing a width between the pair of parallel lines that have been created in the reference parallel line creation step with a maximum separation threshold of the welding electrodes, and determining whether the maximum separation threshold is larger than the width between the parallel lines, and a pull-out direction determination step of determining a direction of the parallel lines as a pull-out direction, in a case where the maximum separation threshold is larger than the width between the parallel lines in the first determination step.