54 results about "Multiple hypothesis" patented technology

In language evaluation systems, user expressions are often evaluated by speech recognizers and language parsers, and among several possible translations, a highest-probability translation is selected and added to a dialogue sequence. However, such systems may exhibit inadequacies by discarding alternative translations that may initially exhibit a lower probability, but that may have a higher probability when evaluated in the full context of the dialogue, including subsequent expressions. Presented herein are techniques for communicating with a user by formulating a dialogue hypothesis set identifying hypothesis probabilities for a set of dialogue hypotheses, using generative and / or discriminative models, and repeatedly re-ranks the dialogue hypotheses based on subsequent expressions. Additionally, knowledge sources may inform a model-based with a pre-knowledge fetch that facilitates pruning of the hypothesis search space at an early stage, thereby enhancing the accuracy of language parsing while also reducing the latency of the expression evaluation and economizing computing resources.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

A system and method for predicting an inter-layer predicted slice of image data from at least two different reference layers, where the inter-layer predicted slice of image data itself resides on yet another layer, different from either of the two reference layers. At least one coded block from the inter-layer predicted slice of image data is encoded with an indication, indicating to a decoder that the at least one coded block is to be inter-layer multi-predicted from the at least two different reference layers. Identifications and corresponding prediction weights of the at least two different reference layers are also signaled to the decoder either in the coded block itself, or in the inter-layer predicted slice of image data.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

A method and process provides structure recognition to a node-link diagram formed by a plurality of digital ink strokes drawn on a canvas of an electronic device by an input device. The method and process include grouping related strokes into multiple hypotheses reflecting structure interpretations of the strokes in a group. Confidence scores are computed for the multiple hypotheses based on local evidence regarding the strokes of the stroke groups and surrounding strokes. Constraints are applied among the hypotheses, and a collectively high scoring assignment of accept / reject values of the hypotheses are determined under the applied constraints. The hypotheses with collectively high scoring assignments are accepted as a representation of the node-link diagram, where structure information is provided to the strokes of the node-link diagram, making the node-link diagram electronically editable.

The navigation solution of a portable device may be enhanced using map information. Sensor data for the portable device may be used to derive navigation solutions at a plurality of epochs over a first period of time. Position information for the device may be estimated at a time subsequent to the first period of time using the navigation solutions. Map information for an area encompassing a current location of the portable device may also be obtained. Multiple hypotheses regarding possible positions of the portable device may be generated using the estimated position information and the map information. By managing and processing the hypotheses, estimated position information for at least one epoch during the first period of time may be updated. An enhanced navigation solution for the at least one epoch may be provided using the updated estimated position information and an uncertainty measure may be derived for the enhanced navigation solution.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

An automated method and system for autocontouring organs and other anatomical structures in CT and other medical images employs one or more contouring techniques, depending on the particular organs or structures to be contoured. In a preferred embodiment, an edge-based technique is employed to contour one or more organs. A multiple hypothesis testing technique can be employed to improve the accuracy of the resulting contour. Independent algorithms can be employed for contouring multiple organ in a given region, such as the male pelvic region. An integration algorithm can be employed to combine the results of the independent algorithms to improve accuracy further.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

The invention discloses a foundation pile foundation model analyzing method and a test device for the same. The method includes: building a dispersing spring support model of to-be-tested pile-earth related actions, building a to-be-tested pile inner force and deformation equation and transmission matrix, and a to-be-tested pile foundation reaction analyzing basic equation, building different founding reaction models and programming a corresponding computerizing program, arranging the foundation pile foundation test device in the field, measuring horizontal displacement and turning angle, on the ground, of the to-be-tested pile, and computerizing and analyzing according to the computerizing program to determine corresponding foundation reaction model parameters. The foundation pile foundation test device comprises a standard beam, a standard pile, a reaction pile and the to-be-tested pile which are disposed on the ground, a jack and a force sensor are connected between the reaction pile and the to-be-tested pile, and a angle measuring device and a displacement measuring device are disposed on the to-be-tested pile. By on-site pile testing, multiple hypothesis models can be built for the foundation pile foundation, related model parameters can be determined, and certain universality is achieved.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

A method and process provides structure recognition to a node-link diagram formed by a plurality of digital ink strokes drawn on a canvas of an electronic device by an input device. The method and process include grouping related strokes into multiple hypotheses reflecting structure interpretations of the strokes in a group. Confidence scores are computed for the multiple hypotheses based on local evidence regarding the strokes of the stroke groups and surrounding strokes. Constraints are applied among the hypotheses, and a collectively high scoring assignment of accept / reject values of the hypotheses are determined under the applied constraints. The hypotheses with collectively high scoring assignments are accepted as a representation of the node-link diagram, where structure information is provided to the strokes of the node-link diagram, making the node-link diagram electronically editable.

A fire control system for a boost phase threat missile includes sensors for generating target-missile representative signals, and a multi-hypothesis track filter, which estimates the states of various target hypotheses. The estimated states are typed to generate hypotheses and their likelihoods. The states, hypotheses and likelihoods are applied to a multihypothesis track filter, and the resulting propagated states are applied to an engagement planner, together with the hypotheses and likelihoods. The engagement planner initializes the interceptor(s). Interceptor guidance uses the initialization and the propagated states and typing information to command the interceptor.

Disclosed a multi-sensor multiple hypotheses testing tracking system. The multi-hypothesis testing system associates measurements from multiple sensors with tracks. Measurements are incorporated using a Kalman Filter and the same filters are used to propagate the trajectories of the tracks.

A video surveillance system uses rule-based reasoning and multiple-hypothesis scoring to detect predefined behaviors based on movement through zone patterns. Trajectory hypothesis spawning allows for trajectory splitting and / or merging and includes local pruning to managed hypothesis growth. Hypotheses are scored based on a number of criteria, illustratively including at least one non-spatial parameter. Connection probabilities computed during the hypothesis spawning process are based on a number of criteria, illustratively including object size. Object detection and probability scoring is illustratively based on object class.

The invention provides a method that enables a search and browse of a document image group through the application of a document structure analysis technique and a character recognition technique as searching / browsing means for paper documents and document images. A highly functional document image search / browse system separates an OCR and a document processing apparatus, adopts as OCR output formats data (reading hypothesis data) holding multiple hypotheses of character line extraction, character segmentation and character recognition, and document structure data having ruled line information, frame information, character line information, browse attribute information and the like about a document image, and provides a function of important keyword extraction and document search from typed and handwritten character strings using OCR-added data, and of document display intended by a browser using the document structure data.

The invention provides a radar data association method based on a simplified multi-hypothesis algorithm. In view of problems of combination explosion and calculation amount exponential rising caused by a delay decision making mechanism of the traditional multi-hypothesis algorithm in a multi-target tracking system, a likelihood ratio scoring function and a linear distribution (LAP) pruning method are introduced, a track association logarithm likelihood ratio is used for simplifying hypothesis association probability calculation, M optimal hypothesis at the current moment is obtained through LAP algorithm track level pruning, through further global level hypothesis pruning, an optimal hypothesis matching sequence is obtained quickly as an effective association measurement point, and the purpose of reducing the calculation amount finally is realized. The method particularly comprises three main steps of track scoring calculation based on the likelihood ratio function, LAP-based track level M-optimal hypothesis pruning and LAP-based global level optimal hypothesis generation. By using the method provided by the invention, on the basis of ensuring the association accuracy, the calculation amount during a multiple hypothesis algorithm data association process can be reduced to a large degree.

The invention discloses a video coding and decoding method based on multiple description CS measurement value. The coding method comprises steps: a video image sequence is divided into a key frame and a CS frame; block-based measurement, quantization and entropy coding are adopted for the key frame; a multiple description idea is adopted for the CS frame to describe a block-based measurement value in two ways; and quantization and entropy coding are separately carried out. The decoding method comprises steps: after key frame data are received, inverse quantization is firstly carried out on entropy decoding to obtain a measurement value, the key frame block is initially reconstructed, and multiple hypothesis prediction and reconstruction are carried out inside the frame; after CS frame multiple description code stream is received, inverse quantization is carried out on entropy decoding; a measurement value based on multiple description is generated; and CS frame multiple description adaptive inter-frame multiple hypothesis reconstruction is carried out. In the condition of not increasing complexity of the coding end, the reconstruction quality and robustness of a video image are improved.

A method and apparatus for decoding encoded data bits of a wireless communication transmission are provided. A set of a-priori bit values corresponding to known bit values of the encoded data bits may be generated. Decoding paths that correspond to decoded data bits that are inconsistent with the a-priori bit values may be removed from the possible decoding paths to consider, and decoding the encoded data bits by selecting a decoding path from remaining decoding paths of the possible decoding paths that were not removed. Multiple hypotheses, each corresponding to a different set of a-prior bit values may be evaluated, with the decoded data for a hypothesis selected based on the evaluation output for further processing.

A method and system for correcting butting artifacts in x-ray images is disclosed. In order to correct a butting artifact in an x-ray image, a butting artifact region in the x-ray image is normalized. Multiple intensity shift estimators are calculated for each pixel of each line of the butting artifact. Confidence intervals are calculated for each intensity shift estimator. A multiple hypothesis hidden Markov model (MH-HMM) is formulated based on the intensity shift operators and confidence measures subject to a smoothness constraint, and the MH-HMM is solved to determine intensity shift values for each pixel. A corrected image is generated by adjusting the intensity of each pixel of the butting artifact based on the intensity shift value for that pixel.

The invention discloses a position estimation optimization method for advanced receiver autonomous integrity monitoring. The invention puts forwards an optimization method for position estimation in allusion to the availability of advanced receiver autonomous integrity monitoring. When a positioning result obtained according to the traditional method in ARAIM (Advanced Receiver Autonomous Integrity Monitoring) does not satisfy an available criterion, a constellation weighted positioning result is adopted to replace the original positioning result within an allowable range of the positioning precision, and the protection level and effective monitoring threshold are recalculated. The method can reduce the difference in positioning precision of different constellations and improve the availability of ARAIM. The method has the characteristics of low computation complexity and good compatibility. The method is suitable for autonomous integrity monitoring of satellite navigation receivers, and the same idea is applicable to other signal systems adopting multiple hypothesis solution separation.

Slotted Aloha-NOMA (SAN) protocol is an uncoordinated, non-orthogonal, random access protocol that exploits the simplicity of SA (Slotted Aloha) and the superior throughput of non-orthogonal multiple access (NOMA) and its ability to resolve collisions via use of successive interference cancellation (SIC) receiver. In SAN protocol, the SIC receiver at the IoT gateway adaptively learns the number of active devices (which is not known a priori) using multiple hypothesis testing in order to successfully distinguish between signals transmitted from different IoT devices.

The invention belongs to the field of computer vision, discloses a multiple hypothesis combined view selection based dense three-dimensional reconstruction method and provides an asymmetric chessboardgrid transmission and multiple hypothesis combined view selection based quick and accurate dense three-dimensional reconstruction method. On the basis of matching cost of neighborhood pixels, asymmetric chessboard grid transmission enables low-cost hypothesis to be preferentially transmitted to a further area to increase algorithm convergence rate. Further, multiple hypotheses are selected by themultiple hypothesis combined view selection method through asymmetric chessboard grid transmission to construct a cost matrix in neighborhood views to extract an appropriate view aggregation set, sothat high accuracy in aggregation cost representation and optimal hypothesis selection is realized. By the method, quickness in algorithm convergence is realized on the premise of making full use of parallel computing power, and accurate robust dense reconstruction results can be obtained.

The invention belongs to the technical field of image encoding and decoding and discloses a differential image compression perception reconfiguration method based on multi-hypothesis weighting and an intelligent terminal. A cross-block-based block compression sensing process is used to sample an original image to obtain a measurement value. A non-local means fully differential iterative reconstruction algorithm is used to carry out iterative reconfiguration processing on the obtained measurement value, and an initial image reconstruction value is obtained. The multiple-hypothesis set acquisition processing is carried out on a current reconstruction value, and the optimized filtering processing is carried out on an obtained multi-hypothesis set to eliminate an inferior hypothesis. The optimized multi-hypothesis set is processed by using a weight estimation model based on an elastic network, a multi-hypothesis weight matrix is obtained, the weighted summation processing is carried out on the multiple hypotheses to obtain side information, and a more accurate iteration initial value is provided for subsequent iterations. According to the method and the intelligent terminal, the spatial correlation of the image is effectively used, the multi-hypothesis weighting processing is used to effectively relieve an over-smoothing problem of a past reconstruction algorithm, and the image reconstruction quality is greatly improved.