Future Wei Technologies, Inc.

University of Florida Research Foundation Inc

Vid Scale, Inc.

Dapeng Oliver Wu

Hong Heather Yu

Jun Tian

Taoran Lu

Yu Huang

Kairan Sun

Qiuyuan Huang

Zheng Yuan

Lei Yang

Xin Li

Bing Han

Qian Chen

Pratik Prabhanjan Brahma

Huazi Zhang

Yun Zhu

Jiade Li

Hong Heather Wu

Zhifeng Chen

Chaoyue Sun

Various embodiments of the present disclosure provide systems and methods for generally generating and transforming physical quantum representations of multi-dimensional tensor data objects. Specifically, various embodiments enable the rapid and efficient generation of a physical quantum representation representing the Fourier transform of a multi-dimensional tensor data object based at least in part on manipulating another physical quantum representation of the multi-dimensional tensor data object itself via quantum manipulation operations. Information may be extracted from the generated physical quantum representation to determine the Fourier transform of the multi-dimensional tensor data object. Accordingly, various embodiments may comprise quantum manipulation operations for a tensor-form quantum Fourier transform (TQFT) for a multi-dimensional tensor data object. Various embodiments for the TQFT are advantageously comprehensive, versatile, and applicable to any quantum representation form for a multi-dimensional tensor data object. The TQFT may be performed in any quantum computing system and/or simulated quantum computing system.

Physical transformations for multi-dimensional data quantum representations

Systems and methods for a computer-based visual recognition based upon a spatially forked deep learning architecture, including unification of deep learning and reasoning. A method can include a primary learner with an adjacent structured memory bank that can not only predict the output from a given input, but also relate the input to all past memorized instances and help in creative understanding.

Deep learning for characterizing unseen categories

A network system for increasing data throughput and decreasing transmission delay along a data link from a source node to a sink node via a relay node is provided. The network system may include a first node configured to encode a second multitude of data packets based on an estimated rank distribution expressing a quality of the data link and transmit the encoded second multitude of data packets. The network system may also include at least a second node configured to estimate the rank distribution based on a first multitude of data packets received from the first node prior to receiving at least one of the encoded second multitude of data packets, transmit the estimated rank distribution to the first node, and regenerate the second multitude of data packets if receiving a sufficient quantity of the encoded second multitude of data packets.

Joint fountain coding and network coding for loss-tolerant information spreading

Network protocols and methods of transmitting data over wireless networks, including wireless mesh networks, are provided. A method of data transmission in wireless networks can include sending packets from one or more terminal nodes to one or more relay nodes, recoding the packets in the one or more relay nodes, and transmitting the recoded packets from the one or more relay nodes to the one or more terminal nodes. The relay nodes can linearly recombine the received packets before retransmission. The terminal nodes can utilize the linearly recombined terminal nodes to recover packets by solving a set of linear equations using their original packets.

Joint fountain code and network coding for multiple-source-multiple-destination wireless communication

A network system for increasing data throughput and decreasing transmission delay from a source node to a sink node via a relay node. The network system may comprise a source node configured to encode a plurality of data packets using rateless coding and transmit the plurality of data packets; at least one relay node configured to receive at least one of the plurality of data packets from the source node, and if the at least one relay node has received a sufficient quantity of the plurality of data packets, regenerate, re-encode, and relay the plurality of data packets; and a sink node configured to receive one or more of the plurality of data packets from the at least one relay node, and if the sink node has received the sufficient quantity of the plurality of data packets, regenerate and decode the plurality of data packets.

A wireless system can include a plurality of directional antennas and a scheduler scheduling multiple links between the plurality of directional antennas. The scheduler can schedule the multiple links concurrently in the same time slot based on Regret Benefit Ratio (RBR). The RBR is a single indicator combining contention for one of the directional antennas and Quality of Service (QoS). The scheduler can introduce a contention graph for the directional antennas that capture a global knowledge of interference.

Regret benefit ratio link scheduler for wireless backhaul with directional antennas

In an embodiment, a method of creating a skimming preview of a video includes electronically receiving a plurality of video shots, analyzing each frame in a video shot from the plurality of video shots, where analyzing includes determining a saliency of each frame of the video shot. The method also includes determining a key frame of the video shot based on the saliency of each frame the video shot, extracting visual features from the key frame, performing shot clustering of the plurality of video shots to determine concept patterns based on the visual features, and generating a reconstruction reference tree based on the shot clustering. The reconstruction reference tree includes video shots categorized according to each concept pattern.

Video skimming methods and systems

A system and method for image and video compression using compressive sensing is provided. An embodiment method for compressing an image having a plurality of image blocks includes selecting an image block from the plurality of image blocks to compress, computing a sum of sinusoidal signals at different frequencies and amplitudes representation for the selected image block, quantizing the amplitudes of the sinusoidal signals at different frequencies, and saving the quantized amplitudes as a first compressed image block. The method also includes truncating the quantized amplitudes, thereby producing truncated quantized amplitudes, saving the truncated quantized amplitudes as a second compressed image block, and selecting either the first compressed image block or the second compressed image block as a final compressed image block. The selecting is based on a measure of the quality of the first compressed image block and the second compressed image block.

System and method for compressing images and video

A system and method for securing media content are provided. A method for distributing security media content includes embedding a marking payload into media content, thereby producing embedded media content, applying a transform to the embedded media content, thereby producing secured media content, and distributing the secured media content. The marking payload is embedded based on perceptual information regarding the media content.

System and method for securing media content

A system and method for retargeting video sequences are provided. A method for retargeting a video includes a plurality of frames includes determining saliency information for the plurality of frames, determining a cost metric for the video, and retargeting the video based on the cost metric to produce a retargeted video. The cost metric considers loss due to cropping, scaling, temporal factors, and spatial factors. The retargeting makes use of a crop window for each frame in the plurality of frames.

System and method for retargeting video sequences

In accordance with an embodiment, a method of comparing a first group of frames to a second group of frames includes electronically receiving the first group of frames, selecting a group of frames from the first group of frames as a first key frame set, calculating a hash distance between an image hash for each frame in the first key frame set to an image hash of each frame of a second key frame set taken from second group of frames, and choosing frames in the first group of frames with a minimum hash distances to respective reference frames to form a series of minimum hash distances.

Video signature based on image hashing and shot detection

A system and method for image and video compression and decompression using compressive sensing is provided. A method for decompressing a compressed image, where the compressed image having a plurality of compressed image blocks, and the method is performed on a processor, includes selecting a compressed image block, entropy decoding the selected compressed image block, and recovering an image block corresponding to the decoded selected compressed image block using compressive sensing recovery.

System and method for compressing and decompressing images and video

In accordance with an embodiment, a method of authenticating images includes electronically receiving an anchor image and a query image, performing a feature point extraction of an anchor image, and performing a feature point extraction of a query image. The method also includes clustering feature points of the anchor image and feature points of the query image, where clustering includes determining matching feature points, determining outlier feature points, and excluding outlier feature points. Whether the anchor image is similar to the query image is determined based on a distance between the feature points of the anchor image and the feature points of the query image. If the anchor image is similar to the query image, possible tampered areas of the query image based on the outlier feature points are identified.

Growing community of inventors

Gainesville, FL, United States of America

Dapeng Oliver Wu