Abstract: Systems and methods are disclosed for dynamically smoothing images based on network conditions to adjust a bitrate needed to transmit the images. Content in the images is smoothed to reduce the quantity of bits needed to encode each image. Filtering the images modifies regions including content having a high frequency of pixel variation, reducing the frequency, so the pixel colors in the region appear “smoothed” or homogeneous. In other words, a region of an image showing a grassy lawn has a high frequency of variation from pixel to pixel resulting from the fine detail of separate blades of grass that may be similar in color, but not homogeneous. Encoding the region as a single shade of green (or multi-pixel regions of different shades of green) enables a viewer to recognize it as a grassy lawn while greatly reducing the number of bits needed to represent the region.

Abstract: Apparatuses, systems, and techniques for isolating the performance of a quality-of-service (QoS) policy for improved data streaming systems and applications. In at least one embodiment, a metric is determined for a QoS policy used to provide an application session based on a value of at least one characteristic of the application session that reflects an impact of one or more external conditions beyond the control of the QoS policy.

Abstract: In various examples, a frame may be encoded as multiple sub-frames. For example, data particularly relevant to conveying visual motion between frames may be encoded in a first sub-frame(s) with remaining data being encoded in a second sub-frame(s). Other information may be included in the first sub-frame(s), such as high entropy data. The high entropy data may be estimated using quantization and dequantization of macroblocks. Packet pacing may be applied at least between the encoded sub-frames. As the first sub-frame(s) may include the most important information for frame updates at the client device, if the second sub-frame(s) is not received and/or displayed the first sub-frame may be displayed providing high quality results. More error correction may be used for the first sub-frame than the second sub-frame to increase the likelihood that the first sub-frame is received at a client device.

Abstract: In various examples, a deep neural network (DNN) based pre-filter for content streaming applications is used to dynamically adapt scene entropy (e.g., complexity) in response to changing network or system conditions of an end-user device. For example, where network and/or system performance issues or degradation are identified, the DNN may be implemented as a frame pre-filter to reduce the complexity or entropy of the frame prior to streaming—thereby allowing the frame to be streamed at a reduced bit rate without requiring a change in resolution. The DNN-based pre-filter may be tuned to maintain image detail along object, boundary, and/or surface edges such that scene navigation—such as by a user participating in an instance of an application—may be easier and more natural to the user.

Abstract: In various examples, machine learning of encoding parameter values for a network is performed using a video encoder. Feedback associated with streaming video encoded by a video encoder over a network may be applied to an MLM(s). Using such feedback, the MLM(s) may predict a value(s) of an encoding parameter(s). The video encoder may then use the value to encode subsequent video data for the streaming. By using the video encoder in training, the MLM(s) may learn based on actual encoded parameter values of the video encoder. The MLM(s) may be trained via reinforcement learning based on video encoded by the video encoder. A rewards metric(s) may be used to train the MLM(s) using data generated or applied to the physical network in which the MLM(s) is to be deployed and/or a simulation thereof. Penalty metric(s) (e.g., the quantity of dropped frames) may also be used to train the MLM(s).

Abstract: Systems and methods are disclosed for dynamically smoothing images based on network conditions to adjust a bitrate needed to transmit the images. Content in the images is smoothed to reduce the quantity of bits needed to encode each image. Filtering the images modifies regions including content having a high frequency of pixel variation, reducing the frequency, so the pixel colors in the region appear “smoothed” or homogeneous. In other words, a region of an image showing a grassy lawn has a high frequency of variation from pixel to pixel resulting from the fine detail of separate blades of grass that may be similar in color, but not homogeneous. Encoding the region as a single shade of green (or multi-pixel regions of different shades of green) enables a viewer to recognize it as a grassy lawn while greatly reducing the number of bits needed to represent the region.