Patent
Method for analysing media content
Reads0
Chats0
TLDR
In this article, the authors proposed a method for analyzing media content, which comprises receiving media content objects by a feature extractor for extracting a plurality of feature maps from said media contents objects, processing the plurality of maps in a bidirectional LSTM neural network, where the LSTMs are aligned along different directions of the feature maps to produce low resolution feature maps, upsampling the low-resolution feature maps according to the size of received media content and assigning each pixel of the upsampled feature maps with a label of maximum likelihood.Abstract:
The invention relates to a method, an apparatus and a computer program product for analyzing media content. The method comprises receiving media content objects by a feature extractor for extracting a plurality of feature maps from said media content objects; processing the plurality of feature maps in a bidirectional Long-Short Term memory neural network, where the bidirectional Long-Short Term memory neural network is aligned along different directions of the feature maps to produce low resolution feature maps; upsampling the low resolution feature maps to the size of received media content; and assigning each pixel of the upsampled feature maps with a label of maximum likelihood for segmenting objects from the upsampled feature maps.read more
Citations
More filters
Patent
Spatial and temporal information for semantic segmentation
TL;DR: Systems and methods for segmenting an image using a convolutional neural network are described herein, and one or more LSTM layers may be combined in such a way that the CNN is predictive, providing predicted output of segmented images.
Patent
Video frame rate up-conversion method and system based on convolutional neural network
TL;DR: In this paper, a video frame rate up-conversion method based on a convolutional neural network is proposed, where the initial video is divided into multiple image blocks of which each comprises two consecutive frame images.
Patent
Basketball game semantic event recognition method based on domain knowledge and deep multi-stage feature
TL;DR: Wang et al. as discussed by the authors proposed a basketball game semantic event recognition method based on domain knowledge and deep multi-stage features to realize automatic recognition of semantic events in basketball games, which includes the steps of dividing basketball semantic events into an event preparation stage, an event occurrence stage and an event follow-up stage.
Patent
A target detection method and system
TL;DR: In this article, a target detection method was proposed for computer videos, which comprises the following steps of: extracting image features of the input image by adopting a convolutional neural network; obtaining a plurality of feature maps with different spatial resolutions of an image, adjusting the spatial resolution of the feature maps to be consistent, splicing the plurality of features into an integral feature map, processing the integral feature maps, and enabling the detection information to comprise target position information and target category information.
Patent
Method of board lumber grading using deep learning techniques
Revathy Priyanga Narasimhan,Patrick S. Freeman,Hayden Michael Aronson,Kevin Johnsrude,Chris Mosbrucker,Dan Robin,Ryan T. Shear,Joseph H. Weintraub,Eric N. Mortensen +8 more
TL;DR: In this article, a method of board lumber grading is performed in an industrial environment on a machine learning framework configured as an interface to a deep convolutional network that is trained end-to-end, pixels-topixels on semantic segmentation.
References
More filters
Patent
Three-dimensional (3D) convolution with 3D batch normalization
TL;DR: In this paper, a 3D deep convolutional neural network architecture (DCNNA) equipped with so-called subnetwork modules which perform dimensionality reduction operations on 3D radiological volume before the volume is subjected to computationally expensive operations is presented.
Patent
Automatic defect classification without sampling and feature selection
TL;DR: In this article, a defect classification system for semiconductor wafer defect detection using a deep-architecture neural network (DANN) is presented. But, the system is limited to the first convolutional layer of the neural network and does not include a subsampling layer.