Nvidia Corporation

Texas Instruments Corporation

Honeywell International Inc.

Keshava Prasad

Hassane S Azar

Vinayak Jayaram Pore

Aaron Daniels

Vijayshekhar Araganji

Pavan Venkata Shastry

John Jerred

Subramanian Esakki

Diksha Garg

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.

Entropy-based pre-filtering using neural networks for streaming applications

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.

Dynamic image smoothing based on network conditions

Embodiments generally relate to linear variable displacement transformer (LVDT) position sensors. The position sensor comprises a bobbin, a moveable core, a primary coil of wire wound around the bobbin, and two secondary coils of wire wound around the bobbin about the primary coil using a fractional winding technique. For example, the winding length of the bobbin may be separated into three consecutive parts. Generally, the primary coil may be wound around the entire winding length of the bobbin. The first secondary coil may be wound around the first and second parts of the winding length. The second secondary coil may be wound around the second and third parts of the winding length. Additionally, the first secondary coil and the second secondary coil may overlap over the second part of the winding length.

Linear variable displacement transformer (LVDT) with improved sensitivity and linearity using fractional winding technique

This invention is a method of memory saving in compressed video decoding. For each group of pictures the method determines whether less than the maximum number N reference frames are needed for decoding. If so, then a memory is configuring for the needed number M reference frames less than N. If not, the memory is configured for N reference frames. The group of pictures is decoding the group of pictures using the configured number of reference frames.

Method for low memory footprint compressed video decoding

Growing community of inventors

Bengaluru, India

Keshava Prasad