Abstract: An encoder system trains a compression model that includes an autoencoder model and a frame extractor model. The encoding portion of the autoencoder is coupled to receive a set of target frames and a previous state tensor for the set of target frames and generate compressed code. The decoding portion of the autoencoder is coupled to receive the compressed code and the previous state tensor for the set of frames and generate a next state tensor for the set of target frames. The frame extractor model is coupled to receive the next state tensor and generate a set of reconstructed frames that correspond to the set of target frames by performing one or more operations on the state tensor. The state tensor for the set of frames includes information from frames of the video that can be used by the frame extractor to generate reconstructed frames.

Abstract: A deep learning based compression (DLBC) system applies trained models to compress binary code of an input image to a target codelength. For a set of binary codes representing the quantized coefficents of an input image, the DLBC system applies a first model that is trained to predict feature probabilities based on the context of each bit of the binary codes. The DLBC system compresses the binary code via adaptive arithmetic coding based on the determined probability of each bit. The compressed binary code represents a balance between a reconstruction quality of a reconstruction of the input image and a target compression ratio of the compressed binary code.

Abstract: An encoder system trains a compression model that includes an autoencoder model and a frame extractor model. The encoding portion of the autoencoder is coupled to receive a set of target frames and a previous state tensor for the set of target frames and generate compressed code. The decoding portion of the autoencoder is coupled to receive the compressed code and the previous state tensor for the set of frames and generate a next state tensor for the set of target frames. The frame extractor model is coupled to receive the next state tensor and generate a set of reconstructed frames that correspond to the set of target frames by performing one or more operations on the state tensor. The state tensor for the set of frames includes information from frames of the video that can be used by the frame extractor to generate reconstructed frames.

Abstract: A compression system includes an encoder and a decoder. The encoder can be deployed by a sender system to encode a tensor for transmission to a receiver system, and the decoder can be deployed by the receiver system to decode and reconstruct the encoded tensor. The encoder receives a tensor for compression. The encoder also receives a quantization mask and probability data associated with the tensor. Each element of the tensor is quantized using an alphabet size allocated to that element by the quantization mask data. The encoder compresses the tensor by entropy coding each element using the probability data and alphabet size associated with the element. The decoder receives the quantization mask data, the probability data, and the compressed tensor data. The quantization mask and probabilities are used to entropy decode and subsequently reconstruct the tensor.

Abstract: Systems, methods, and non-transitory computer-readable media can identify a set of regions corresponding to a geographical area. A collection of training images can be acquired. Each training image in the collection can be associated with one or more respective recognized objects and with a respective region in the set of regions. Histogram metrics for a plurality of object categories within each region in the set of regions can be determined based at least in part on the collection of training images. A neural network can be developed based at least in part on the histogram metrics for the plurality of object categories within each region in the set of regions and on the collection of training images.

Abstract: An enhanced encoder system generates residual bitstreams representing additional image information that can be used by an image enhancement system to improve a low quality image. The enhanced encoder system upsamples a low quality image and compares the upsampled image to a true high quality image to determine image inaccuracies that arise due to the upsampling process. The enhanced encoder system encodes the information describing the image inaccuracies using a trained encoder model as the residual bitstream. The image enhancement system upsamples the same low quality image to obtain a prediction of a high quality image that can include image inaccuracies. Given the residual bitstream, the image enhancement system decodes the residual bitstream using a trained decoder model and uses the additional image information to improve the predicted high quality image. The image enhancement system can provide an improved, high quality image for display.

Abstract: An encoder system trains a compression model that includes an autoencoder model and a frame extractor model. The encoding portion of the autoencoder is coupled to receive a set of target frames and a previous state tensor for the set of target frames and generate compressed code. The decoding portion of the autoencoder is coupled to receive the compressed code and the previous state tensor for the set of frames and generate a next state tensor for the set of target frames. The frame extractor model is coupled to receive the next state tensor and generate a set of reconstructed frames that correspond to the set of target frames by performing one or more operations on the state tensor. The state tensor for the set of frames includes information from frames of the video that can be used by the frame extractor to generate reconstructed frames.

Abstract: An autoencoder is configured to encode content at different quality levels. The autoencoder includes an encoding system and a decoding system with neural network layers forming an encoder network and a decoder network. The encoder network and decoder network are configured to include branching paths through the networks that include different subnetworks. During deployment, content is provided to the encoding system with a quality signal indicating a quality at which the content can be reconstructed. The quality signal determines which of the paths through the encoder network are activated for encoding the content into one or more tensors, which are compressed into a bitstream and later used by the decoding system to reconstruct the content. The autoencoder is trained by randomly or systematically selecting different combinations of tensors to use to encode content and backpropagating error values from loss functions through the network paths associated with the selected tensors.

Abstract: A sample set of images is received. Each image in the sample set may be associated with one or more social cues. Correlation of each image in the sample set with an image class is scored based on the one or more social cues associated with the image. Based on the scoring, a training set of images to train a classifier is determined from the sample set. In an embodiment, an extent to which an evaluation set of images correlates with the image class is determined. The determination may comprise ranking a top scoring subset of the evaluation set of images.

Abstract: A compression system trains a machine-learned encoder and decoder. The encoder can be deployed by a sender system to encode content for transmission to a receiver system, and the decoder can be deployed by the receiver system to decode the encoded content and reconstruct the original content. The encoder receives content and generates a tensor as a compact representation of the content. The content may be, for example, images, videos, or text. The decoder receives a tensor and generates a reconstructed version of the content. In one embodiment, the compression system trains one or more encoding components such that the encoder can adaptively encode different degrees of information for regions in the content that are associated with characteristic objects, such as human faces, texts, or buildings.

Abstract: A deep learning based compression (DLBC) system generates a progressive representation of the encoded input image such that a client device that requires the encoded input image at a particular target bitrate can readily be transmitted the appropriately encoded data. More specifically, the DLBC system computes a representation that includes channels and bitplanes that are ordered based on importance. For a given target rate, the DLBC system truncates the representation according to a trained zero mask to generate the progressive representation. Transmitting a first portion of the progressive representation enables a client device with the lowest target bitrate to appropriately playback the content. Each subsequent portion of the progressive representation allows the client device to playback the content with improved quality.

Abstract: Various embodiments of the present disclosure include systems, methods, and non-transitory computer storage media configured to identify a set of training content items, each of the set of training content items comprising video content. A category may be assigned to each of the set of training content items. A plurality of variations may be provided to the each of the set of training content items. A first content recognition module may be trained in an unsupervised process to associate the plurality of variations of the each of the set of training content items with the category assigned to the each of the set of training content items. A classification layer may be generated based on the training the first content recognition module in the unsupervised process. A second content recognition module may be trained in a supervised process based on the classification layer.

Abstract: Systems, methods, and non-transitory computer readable media are configured to generate a hash value of an evaluation content item. Hash values of a plurality of content items associated with location information are generated. A pairwise distance value between the hash value of the evaluation content item and a hash value of each content item of the plurality of content items assigned to a group of a plurality of groups is determined. A score for each group of the plurality of groups is generated based on a combination of distance values for the group. At least one location associated with the evaluation content item is determined based on scores for the plurality of groups.

Abstract: A deep learning based compression (DLBC) system generates a progressive representation of the encoded input image such that a client device that requires the encoded input image at a particular target bitrate can readily be transmitted the appropriately encoded data. More specifically, the DLBC system computes a representation that includes channels and bitplanes that are ordered based on importance. For a given target rate, the DLBC system truncates the representation according to a trained zero mask to generate the progressive representation. Transmitting a first portion of the progressive representation enables a client device with the lowest target bitrate to appropriately playback the content. Each subsequent portion of the progressive representation allows the client device to playback the content with improved quality.

Abstract: Systems, methods, and non-transitory computer-readable media can acquire video content for which video feature descriptors are to be determined. The video content can be processed based at least in part on a convolutional neural network including a set of two-dimensional convolutional layers and a set of three-dimensional convolutional layers. One or more outputs can be generated from the convolutional neural network. A plurality of video feature descriptors for the video content can be determined based at least in part on the one or more outputs from the convolutional neural network.

Abstract: A sample set of images is received. Each image in the sample set may be associated with one or more social cues. Correlation of each image in the sample set with an image class is scored based on the one or more social cues associated with the image. Based on the scoring, a training set of images to train a classifier is determined from the sample set. In an embodiment, an extent to which an evaluation set of images correlates with the image class is determined. The determination may comprise ranking a top scoring subset of the evaluation set of images.

Abstract: The techniques introduced here include a system and method for transcoding multimedia content based on the results of content analysis. The determination of specific transcoding parameters, used for transcoding multimedia content, can be performed by utilizing the results of content analysis of the multimedia content. One of the results of the content analysis is the determination of image type of any images included in the multimedia content. The content analysis uses one or more of several techniques, including analyzing content metadata, examining colors of contiguous pixels in the content, using histogram analysis, using compression distortion analysis, analyzing image edges, or examining user provided inputs. Transcoding the multimedia content can include adapting the content to the constraints in delivery and display, processing and storage of user computing devices.

Abstract: A deep learning based compression (DLBC) system applies trained models to compress binary code of an input image to a target codelength. For a set of binary codes representing the quantized coefficents of an input image, the DLBC system applies a first model that is trained to predict feature probabilities based on the context of each bit of the binary codes. The DLBC system compresses the binary code via adaptive arithmetic coding based on the determined probability of each bit. The compressed binary code represents a balance between a reconstruction quality of a reconstruction of the input image and a target compression ratio of the compressed binary code.

Abstract: A deep learning based compression (DLBC) system trains multiple models that, when deployed, generates a compressed binary encoding of an input image that achieves a reconstruction quality and a target compression ratio. The applied models effectively identifies structures of an input image, quantizes the input image to a target bit precision, and compresses the binary code of the input image via adaptive arithmetic coding to a target codelength. During training, the DLBC system reconstructs the input image from the compressed binary encoding and determines the loss in quality from the encoding process. Thus, the models can be continually trained to, when applied to an input image, minimize the loss in reconstruction quality that arises due to the encoding process while also achieving the target compression ratio.

Abstract: The compression system trains a machine-learned encoder and decoder through an autoencoder architecture. The encoder can be deployed by a sender system to encode content for transmission to a receiver system, and the decoder can be deployed by the receiver system to decode the encoded content and reconstruct the original content. The encoder is coupled to receive content and output a tensor as a compact representation of the content. The content may be, for example, images, videos, or text. The decoder is coupled to receive a tensor representing content and output a reconstructed version of the content. The compression system trains the autoencoder with a discriminator to reduce compression artifacts in the reconstructed content. The discriminator is coupled to receive one or more input content, and output a discrimination prediction that discriminates whether the input content is the original or reconstructed version of the content.