Ben Group, Inc.

Branded Entertainment Network, Inc.

Schubert R Carvalho

Tyler Folkman

Richard Ray Butler

Nicolas M Bertagnolli

This invention classifies an action that appears in a video clip by receiving a video clip for analysis, applying a convolutional neural network mechanism (CNN) to the frames in the clip to generate a 4D embedding tensor for each frame in the clip, applying a multi-resolution convolutional neural network mechanism (CNN) to each of the frames in the clip to generate a sequence of reduced resolution blocks, computing a kinematic attention weight that estimates the amount of motion in the block, applying the attention weights to the embedding tensors for each frame in a clip, to generate a weighted embedding tensor, or context, that represents all the frames in the clip, at the resolution, combining the contexts across all resolutions to generate a multi-resolution context, performing a 3D pooling to obtain a 1D feature vector and classifying a primary action of the video clip based on the feature vector.

Multi-resolution attention network for video action recognition

This invention classifies actions performed within a video clip, by receiving a video clip for analysis, where the video clip comprises a time sequence of video frames, applies a bottleneck attention mechanism to the frames in the clip to generate a reduced sequence of key-frames, applies a 2 dimensional (2D) convolutional neural network to the sequence of keyframes to obtain a 3D embedding tensor for each keyframe, applies a multi-headed attention mechanism to the 3D embedding tensors to generate a final action context, and apples a classification mechanism to the final action context to obtain a probability for each action class that indicates the likelihood that an action specified by the action class occurred in the video clip.

Growing community of inventors

Orem, UT, United States of America

Schubert R Carvalho