Product
Patsnap Eureka
For R&D, Patsnap Eureka makes reading and utilizing patents & technical documents easy.
Patsnap Eureka AIR
Designed for self-driven R&D workflows. Generate viable solutions, solve complex R&D challenges, empower your innovation with AI.
Patsnap Eureka Materials
Designed for material experts only. Revolutionize your material R&D, from search, analyze, to developing new materials.
Feature
TechResearch
Generate reliable direction feasibility study reports for your R&D in just a few steps.
TechSeek
Discover and master advanced knowledge NOW. Basics, ideas, possibilities, all at once.
TechMind
As an expert in R&D Theories, TechMind can generates customized viable solutions instantly.
TechRisk
Analyze your overall solution with one click, know your potential R&D risks in advance.
TechMonitor
Get weekly tech updates, stay abreast of the latest tech innovations and key insights.
Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
2984 results about "Fine-tuning" patented technology
Efficacy Topic
Property
Owner
Technical Advancement
Application Domain
Technology Topic
Technology Field Word
Patent Country/Region
Patent Type
Patent Status
Application Year
Inventor
In theoretical physics, fine-tuning is the process in which parameters of a model must be adjusted very precisely in order to fit with certain observations. Theories requiring fine-tuning are regarded as problematic in the absence of a known mechanism to explain why the parameters happen to have precisely the observed values that they return. The heuristic rule that parameters in a fundamental physical theory should not be too fine-tuned is called naturalness.
Text to speech synthesis
An input linguistic description is converted into a speech waveform by deriving at least one target unit sequence corresponding to the linguistic description, selecting from a waveform unit database for the target unit sequences a plurality of alternative unit sequences approximating the target unit sequences, concatenating the alternative unit sequences to alternative speech waveforms and presenting the alternative speech waveforms to an operating person and enabling the choice of one of the presented alternative speech waveforms. There are no iterative cycles of manual modification and automatic selection, which enables a fast way of working. The operator does not need knowledge of units, targets, and costs, but chooses from a set of given alternatives. The fine-tuning of TTS prompts therefore becomes accessible to non-experts.
Owner:CERENCE OPERATING CO
Apparatus and methods for securing tissue to bone
The present invention relates to apparatus and methods for securing tissue to bone using a suture anchoring system that provides enhanced tactile feedback and does not require tying a suture knot. In each embodiment, a surgeon can individually tension the free ends of the suture to fine-tune the placement of the tissue with respect to the bone, and then secure the suture without tying a knot. In several embodiments of the present invention, the device may be transformed between locked and unlocked suture states, thereby allowing further fine-tuning of the tension in the suture.
Owner:C2M MEDICAL INC
Wearable electromyography-based controllers for human-computer interface
A “Wearable Electromyography-Based Controller” includes a plurality of Electromyography (EMG) sensors and provides a wired or wireless human-computer interface (HCl) for interacting with computing systems and attached devices via electrical signals generated by specific movement of the user's muscles. Following initial automated self-calibration and positional localization processes, measurement and interpretation of muscle generated electrical signals is accomplished by sampling signals from the EMG sensors of the Wearable Electromyography-Based Controller. In operation, the Wearable Electromyography-Based Controller is donned by the user and placed into a coarsely approximate position on the surface of the user's skin. Automated cues or instructions are then provided to the user for fine-tuning placement of the Wearable Electromyography-Based Controller. Examples of Wearable Electromyography-Based Controllers include articles of manufacture, such as an armband, wristwatch, or article of clothing having a plurality of integrated EMG-based sensor nodes and associated electronics.
Owner:MICROSOFT TECH LICENSING LLC
Wireless broadcast link to remote receiver
InactiveUS6256303B1Increase rangeImprove noise qualityTelevision system detailsFrequency-division multiplex detailsSignal onCarrier signal
A high frequency signal transmission apparatus includes a transmitter for transmitting received signals from a signal source as a modulated signal on a high frequency 900 MHz carrier to a remote receiver. The remote receiver converts the high frequency carrier in at least one conversion step to a lower frequency carrier signal. The lower frequency carrier signal carrying the modulated signal is then converted to another carrier signal and retransmitted as a modulated second signal to another remote receiver capable of demodulating the signal and broadcasting the audio sounds or video images. The oscillators in the transmitter and receiver are selectable to operate at discrete frequencies. Automatic fine tuning control is provided in the receiver to accommodate drift of the oscillator in the receiver. A recorder is coupled to the receiver for temporarily storing a demodulated first signal from the transmitter and then outputting the stored signals to a second transmitter for retransmission to a remote output device.
Owner:CONCERT TECH
Golf club iron
An iron golf club head having a thin (less than 0.12 inches) first section that has an expanded unsupported front face region. The first section including a central portion forming part of a leading edge and wrapping around a sole section of the club, to create an increase the coefficient of restitution of the club head to greater than 0.8. The club head utilizes a rear insert that in addition to providing support for the front face, also allows for the fine tuning of swing weights with no change in geometry or size of the club head. This is accomplished this by the utilization of weight adjustment inserts that impregnate tungsten loaded plastic into sheets of carbon graphite and epoxy. The percentage of tungsten creating a weight range without any size change in the sheets.
Owner:ACUSHNET CO
Periodic sensor panel baseline adjustment
ActiveUS8026904B2Easy to detectEliminate captureTransmission systemsCathode-ray tube indicatorsEngineeringBaselining
A periodic baseline adjustment algorithm is disclosed that can increment or decrement individual baseline sensor output values by one count to provide periodic fine-tuning of the baseline to track temperature drift. A scan of the sensor panel is performed after a dynamic adjustment time interval has passed. Previously computed offset values are subtracted from the sensor output values to normalize them. For any normalized sensor values that are positive, the offset values are incremented by one count. For any normalized sensor values that are negative, the offset values are decremented by one count. The algorithm waits the duration of a dynamic adjustment period before scanning the panel again. Dynamic adjustment intervals allow quick recovery for slight baseline inversions without adapting out far-field or touch objects very quickly. Adjustment intervals can depend on an average of background (non-patch) sensor output values.
Owner:APPLE INC
Location based power control for mobile communications systems
InactiveUS6845246B1Improve efficiencyQuick implementationPower managementTransmission control/equalisingEngineeringMobile communication systems
A method and apparatus for facilitating power control within mobile radio systems. The power control is a function of location of the mobile station. Power levels are mapped to a geographic grid and power levels are set based upon such pre-determined power levels. Alternatively, initial power levels are set based upon the mapped power levels, but are subject to fine tuning through a much reduced feedback loop.
Owner:MICROSOFT TECH LICENSING LLC
Large core holey fibers
Various types of holey fiber provide optical propagation. In various embodiments, for example, a large core holey fiber comprises a cladding region formed by large holes arranged in few layers. The number of layers or rows of holes about the large core can be used to coarse tune the leakage losses of the fundamental and higher modes of a signal, thereby allowing the non-fundamental modes to be substantially eliminated by leakage over a given length of fiber. Fine tuning of leakage losses can be performed by adjusting the hole dimension and / or the hole spacing to yield a desired operation with a desired leakage loss of the fundamental mode. Resulting holely fibers have a large hole dimension and spacing, and thus a large core, when compared to traditional fibers and conventional fibers that propagate a single mode. Other loss mechanisms, such as bend loss and modal spacing can be utilized for selected modes of operation of holey fibers. Other embodiments are also provided.
Owner:IMRA AMERICA
Multi-layer fusion in a convolutional neural network for image classification
ActiveUS20170140253A1Easy to adaptImprove classification accuracyCharacter and pattern recognitionNeural learning methodsFeature extractionTwo step
A method and system for domain adaptation based on multi-layer fusion in a convolutional neural network architecture for feature extraction and a two-step training and fine-tuning scheme. The architecture concatenates features extracted at different depths of the network to form a fully connected layer before the classification step. First, the network is trained with a large set of images from a source domain as a feature extractor. Second, for each new domain (including the source domain), the classification step is fine-tuned with images collected from the corresponding site. The features from different depths are concatenated with and fine-tuned with weights adjusted for a specific task. The architecture is used for classifying high occupancy vehicle images.
Owner:CONDUENT BUSINESS SERVICES LLC
Wearable electromyography-based controllers for human-computer interface
A “Wearable Electromyography-Based Controller” includes a plurality of Electromyography (EMG) sensors and provides a wired or wireless human-computer interface (HCl) for interacting with computing systems and attached devices via electrical signals generated by specific movement of the user's muscles. Following initial automated self-calibration and positional localization processes, measurement and interpretation of muscle generated electrical signals is accomplished by sampling signals from the EMG sensors of the Wearable Electromyography-Based Controller. In operation, the Wearable Electromyography-Based Controller is donned by the user and placed into a coarsely approximate position on the surface of the user's skin. Automated cues or instructions are then provided to the user for fine-tuning placement of the Wearable Electromyography-Based Controller. Examples of Wearable Electromyography-Based Controllers include articles of manufacture, such as an armband, wristwatch, or article of clothing having a plurality of integrated EMG-based sensor nodes and associated electronics.
Owner:MICROSOFT TECH LICENSING LLC
Compressed low-resolution image restoration method based on combined deep network
The present invention provides a compressed low-resolution image restoration method based on a combined deep network, belonging to the digital image / video signal processing field. The compressed low-resolution image restoration method based on the combined deep network starts from the aspect of the coprocessing of the compression artifact and downsampling factors to complete the restoration of a degraded image with the random combination of the compression artifact and the low resolution; the network provided by the invention comprises 28 convolution layers to establish a leptosomatic network structure, according to the idea of transfer learning, a model trained in advance employs a fine tuning mode to complete the training convergence of a greatly deep network so as to solve the problems of vanishing gradients and gradient explosion; the compressed low-resolution image restoration method completes the setting of the network model parameters through feature visualization, and the relation of the end-to-end learning degeneration feature and the ideal features omits the preprocessing and postprocessing; and finally, three important fusions are completed, namely the fusion of the feature figures with the same size, the fusion of residual images and the fusion of the high-frequency information and the high-frequency initial estimation figure, and the compressed low-resolution image restoration method can solve the super-resolution restoration problem of the low-resolution image with the compression artifact.
Owner:BEIJING UNIV OF TECH
Periodic sensor panel baseline adjustment
ActiveUS20080158182A1Easy to detectEliminate captureInput/output processes for data processingEngineeringDuration time
A periodic baseline adjustment algorithm is disclosed that can increment or decrement individual baseline sensor output values by one count to provide periodic fine-tuning of the baseline to track temperature drift. A scan of the sensor panel is performed after a dynamic adjustment time interval has passed. Previously computed offset values are subtracted from the sensor output values to normalize them. For any normalized sensor values that are positive, the offset values are incremented by one count. For any normalized sensor values that are negative, the offset values are decremented by one count. The algorithm waits the duration of a dynamic adjustment period before scanning the panel again. Dynamic adjustment intervals allow quick recovery for slight baseline inversions without adapting out far-field or touch objects very quickly. Adjustment intervals can depend on an average of background (non-patch) sensor output values.
Owner:APPLE INC
Wearable electromyography-based human-computer interface
ActiveUS9037530B2Input/output for user-computer interactionElectromyographyHuman–machine interfaceSkin surface
A “Wearable Electromyography-Based Controller” includes a plurality of Electromyography (EMG) sensors and provides a wired or wireless human-computer interface (HCI) for interacting with computing systems and attached devices via electrical signals generated by specific movement of the user's muscles. Following initial automated self-calibration and positional localization processes, measurement and interpretation of muscle generated electrical signals is accomplished by sampling signals from the EMG sensors of the Wearable Electromyography-Based Controller. In operation, the Wearable Electromyography-Based Controller is donned by the user and placed into a coarsely approximate position on the surface of the user's skin. Automated cues or instructions are then provided to the user for fine-tuning placement of the Wearable Electromyography-Based Controller. Examples of Wearable Electromyography-Based Controllers include articles of manufacture, such as an armband, wristwatch, or article of clothing having a plurality of integrated EMG-based sensor nodes and associated electronics.
Owner:MICROSOFT TECH LICENSING LLC
Charged particle cancer therapy and patient positioning method and apparatus
The invention comprises a laying, semi-vertical, or seated patient positioning, alignment, and / or control method and apparatus used in conjunction with multi-axis charged particle or proton beam radiation therapy of cancerous tumors. Patient positioning constraints are used to maintain the patient in a treatment position, including one or more of: a seat support, a back support, a head support, an arm support, a knee support, and a foot support. One or more of the positioning constraints are movable and / or under computer control for rapid positioning and / or immobilization of the patient. The system optionally uses an X-ray beam that lies in substantially the same path as a proton beam path of a particle beam cancer therapy system. The generated image is usable for: fine tuning body alignment relative to the proton beam path, to control the proton beam path to accurately and precisely target the tumor, and / or in system verification and validation.
Owner:BALAKIN ANDREY VLADIMIROVICH +1
Attention mechanism-based in-depth learning diabetic retinopathy classification method
ActiveCN108021916AEnhanced feature informationImproving Lesion Diagnosis ResultsImage enhancementImage analysisNerve networkDiabetes retinopathy
The invention discloses an attention mechanism-based in-depth learning diabetic retinopathy classification method comprising the following steps: a series of eye ground images are chosen as original data samples which are then subjected to normalization preprocessing operation, the preprocessed original data samples are divided into a training set and a testing set after being cut, a main neutralnetwork is subjected to parameter initializing and fine tuning operation, images of the training set are input into the main neutral network and then are trained, and a characteristic graph is generated; parameters of the main neutral network are fixed, the images of the training set are adopted for training an attention network, pathology candidate zone degree graphs are output and normalized, anattention graph is obtained, an attention mechanism is obtained after the attention graph is multiplied by the characteristic graph, an obtained result of the attention mechanism is input into the main neutral network, the images of the training set are adopted for training operation, and finally a diabetic retinopathy grade classification model is obtained. Via the method disclosed in the invention, the attention mechanism is introduced, a diabetic retinopathy zone data set is used for training the same, and information characteristics of a retinopathy zone is enhanced while original networkcharacteristics are reserved.
Owner:NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
System and method for providing identity theft security
InactiveUS20110040983A1Reduce the burden onMinimal effortDigital data processing detailsUnauthorized memory use protectionIdentity theftComputerized system
A system and method of providing identity theft security is provided. The system and method utilizes a computer program that identifies, locates, secures, and / or removes from computers, computer systems and / or computer networks personally identifying and / or other sensitive information in different data formats. The computer program utilizes a multi-tiered escalation model of searching / identifying sensitive information. The computer program of the instant invention utilizes a self-learning process for fine-tuning a level of scrutiny for identifying potentially sensitive information.
Owner:TOUCHNET INFORMATION SYST
Power assisted automatic supervised classifier creation tool for semiconductor defects
A method and system that optionally allows a user to view image defects organized by natural groupings based on features of the images. The natural groupings make it easier for the user to organize some or all of the images into classes in a training set of images. A feature vector is extracted from each image in the training set and stored, along with its user-specified class, for use by an automatic classifier software module. The automatic classifier uses the stored feature vectors and classes to automatically classify images not in the training set. If the automatically classified images do not match images manually classified by the user, the user modifies the training set until a better result is obtained from the automatic classifier. The system can provide feedback to an inspection system designed to aid in the setup and fine-tuning of the inspection system.
Owner:KLA TENCOR TECH CORP
Microwave emission mechanism, microwave plasma source and surface wave plasma processing apparatus
ActiveUS20140361684A1Increase probabilityIncrease ratingsElectric discharge tubesAntenna supports/mountingsDielectricElectricity
A microwave emission mechanism includes: a transmission path through which a microwave is transmitted; and an antenna section that emits into a chamber the microwave transmitted through the transmission path. The antenna section includes an antenna having a slot through which the microwave is emitted, a dielectric member through which the microwave emitted from the antenna is transmitted and a closed circuit in which a surface current and a displacement current flow. A surface wave is formed in a surface of the dielectric member. The closed circuit has at least: an inner wall of the slot; and the surface and an inner portion of the dielectric member. When a wavelength of the microwave is λ0, a length of the closed circuit is nλ0±δ, where n is a positive integer and δ is a fine-tuning component including 0.
Owner:TOKYO ELECTRON LTD
Method for analyzing and classifying process data
InactiveUS20100030521A1Facilitate data analysisData is often requiredTesting/monitoring control systemsDigital computer detailsSystems analysisData set
Process data mining system and method. The system analyzes data from complex process plants or systems and operates in open-book and closed-book modes. In closed-book mode, the system monitors incoming data sets against pre-defined clusters of data values and generates reports, indicating whether incoming data is a match or a no-match with the pre-defined clusters. In open-book mode, the system generates initial clusters, without having a-priori knowledge of the component or process, and also creates clusters “on the fly”, thereby fine-tuning the analysis. A knowledge base encompasses a combination of parameters for a particular component. Clusters are defined within the knowledge base, each cluster representing a particular operating condition. The system expands clusters, within pre-defined limits, or creates new clusters, as needed, in order to accommodate incoming data values. Newly created clusters are then named, so as to indicate the particular operating conditions.
Owner:ISAGACITY
Combination self-expandable, balloon-expandable endoluminal device
An endoluminal device, such as a stent or a vena cava filter, comprising at least one superelastic section and at least one plastically deformable section. The superelastic section may comprise, for example, a superelastic grade of nitinol, whereas the plastically deformable section may comprise, for example, gold, platinum, tantalum, titanium, stainless steel, tungsten, a nickel alloy, a cobalt ally, a titanium alloy, or a combination thereof. Each plastically deformable section may merely comprise a constrained portion of the superelastic section comprising a plastically deformable material, such as gold. The device enables deployment by a method comprising introducing the device into a body lumen with the device radially constrained in a first configuration having a first diameter; allowing the device to self-expand into a second configuration having a second diameter less than or equal to a fully-self-expanded diameter; and then optionally “fine-tuning” the device by forcibly expanding the device into a third configuration having a third diameter in a range between the second diameter and less than or equal to a fully-forcibly-expanded diameter. The superelastic and plastically deformable sections may be tubular sections placed end-to-end, such that the plastically deformable section can be conformed to fit a tapered section of a lumen.
Owner:BOSTON SCI SCIMED INC
System and method for providing identity theft security
InactiveUS20100024037A1Reduce the burden onMinimal effortDigital data processing detailsDigital computer detailsComputerized systemIdentity theft
A system and method of providing identity theft security is provided. The system and method utilizes a computer program that identifies, locates, secures, and / or removes from computers, computer systems and / or computer networks personally identifying and / or other sensitive information in different data formats. The computer program utilizes a multi-tiered escalation model of searching / identifying sensitive information. The computer program of the instant invention utilizes a self-learning process for fine-tuning a level of scrutiny for identifying potentially sensitive information.
Owner:TOUCHNET INFORMATION SYST
Patient positioning method and apparatus used in conjunction with a charged particle cancer therapy system
InactiveUS20090314960A1Chemical conversion by chemical reactionX-ray/gamma-ray/particle-irradiation therapySystem verificationFoot supports
The invention comprises a semi-vertical or seated patient positioning, alignment, and / or control method and apparatus used in conjunction with multi-axis charged particle or proton beam radiation therapy of cancerous tumors. Patient positioning constraints are used to maintain the patient in a treatment position, including one or more of: a seat support, a back support, a head support, an arm support, a knee support, and a foot support. One or more of the positioning constraints are movable and / or under computer control for rapid positioning and / or immobilization of the patient. The system optionally uses an X-ray beam that lies in substantially the same path as a proton beam path of a particle beam cancer therapy system. The generated image is usable for: fine tuning body alignment relative to the proton beam path, to control the proton beam path to accurately and precisely target the tumor, and / or in system verification and validation.
Owner:BALAKIN VLADIMIR
Method and system for implicitly resolving pointing ambiguities in human-computer interaction (HCI)
InactiveUS6907581B2Simple design and implementationEasy accessInput/output processes for data processingObject basedAmbiguity
Owner:RAMOT AT TEL AVIV UNIV LTD
Semi-vertical positioning method and apparatus used in conjunction with a charged particle cancer therapy system
The invention comprises a semi-vertical patient positioning, alignment, and / or control method and apparatus used in conjunction with charged particle or proton beam radiation therapy of cancerous tumors. Patient positioning constraints are used to maintain the patient in a treatment position, including one or more of: a seat support, a back support, a head support, an arm support, a knee support, and a foot support. One or more of the positioning constraints are movable and / or under computer control for rapid positioning and / or immobilization of the patient. The system optionally uses an X-ray beam that lies in substantially the same path as a proton beam path of a particle beam cancer therapy system. The generated image is usable for: fine tuning body alignment relative to the proton beam path, to control the proton beam path to accurately and precisely target the tumor, and / or in system verification and validation.
Owner:BALAKIN ANDREY VLADIMIROVICH +1
Method and arrangement for power control
ActiveUS20040043783A1Error prevention/detection by using return channelPower managementCommunications systemPower control
A method for setting a power control level (600) in a wireless communication system. The method includes the steps of obtaining (620) transmission information from a wireless subscriber unit, and modifying a power control level and / or a communication channel format (660) in response to said transmission information. Preferably, the transmission information is re-transmission requests from a wireless subscriber unit, which are transmitted frequently. In this manner, rapid adjustment of power control can be attained utilising an optimal selection of the available communication channel format, followed by a fine-tuning power control operation.
Owner:SONY CORP
Chinese named entity recognition method based on BERT-BiGRU-CRF
ActiveCN110083831AHigh precisionEffectively represent ambiguitySemantic analysisEnergy efficient computingSemantic vectorPattern recognition
The invention discloses a Chinese named entity recognition method based on BERT-BiGRU-CRF. The method comprises three stages: in the first stage, preprocessing mass text corpora, and pre-training a BERT language model; in the second stage, preprocessing the named entity recognition corpus, and encoding the named entity recognition corpus through the trained BERT language model; and at the third stage, inputting the encoded corpus into a BiGRU+CRF model for training, and performing named entity recognition on the to-be-recognized statement by using the trained model. Construction of the Chinesenamed entity recognition method based on BERT-BiGRU-CRF is carried out, semantic representation of characters is enhanced through a BERT pre-training language model, semantic vectors are dynamicallygenerated according to contexts of the characters, and the ambiguity of the characters is effectively represented. Compared with a method based on fine tuning of a language model, the method has the advantages that training parameters are reduced, and the training time is saved.
Owner:WUHAN UNIV
Stationary exercise bicycle
InactiveUS6669603B1Improve reliabilityReduce maintenanceFrictional force resistorsMovement coordination devicesMechanical componentsLife span
The invention is directed to a stationary bicycle which has load-providing components allowing for fine-tuned adjustment of the selected load. The load providing components are made of relatively simple mechanical components so as to increase the reliability, reduce the maintenance and provide a relatively smooth cycling feel. In addition, the invention is directed to a stationary bicycle which is manufactured using specific materials and processes so as to provide for a reliable product having a relatively long life cycle characterized by reliable functioning. Furthermore, the stationary bicycle of the invention comprises a design which allows for ergonomic fine tuning of the relative positioning between the various components thereof so as to provide ergonomic adjustments thereby reducing the risk of injury to the body of the individual and allowing for a more efficient workout.
Owner:MAD DOGG ATHLETICS
Method, system and computer program product for automatic and semi-automatic modificatoin of digital images of faces
ActiveUS20120299945A1Increase perceived beautyImprove accuracyCharacter and pattern recognitionCathode-ray tube indicatorsSemi automaticHand held
The present invention is directed at modifying digital images of faces automatically or semi-automatically. In one aspect, a method of detecting faces in digital images and matching and replacing features within the digital images is provided. Techniques for blending, recoloring, shifting and resizing or portions of digital images are disclosed. In other aspects, methods of virtual “face lifts” are methods of detecting faces within digital image are provided. Advantageously, the detection and localization of faces and facial features, such as the eyes, nose, lips and hair, can be achieved on an automated or semi-automated basis. User feedback and adjustment enables fine tuning of modified images. A variety of systems for matching and replacing features within digital images is also provided, including implementation as a website, through mobile phones, handheld computers, or a kiosk. Related computer program products are also disclosed.
Owner:LOREAL SA
Adaptive extraction and diagnosis method for degree features of mechanical fault through stack-type sparse automatic coding depth neural network
InactiveCN106323636AImprove robustnessRealize intelligent diagnosis of fault levelMachine bearings testingNeural learning methodsAdaptive learningFeature extraction
The invention relates to an adaptive extraction and diagnosis method for degree features of a mechanical fault through a stack-type sparse automatic coding depth neural network, and belongs to the technical field of mechanical equipment state monitoring and reliability evaluation. The method aims at a problem of intelligent diagnosis of the degree of the mechanical fault, and comprises the steps: carrying out the stacking of sparse automatic coding, adding a classification layer, and constructing the stack-type sparse automatic coding depth neural network which integrates the adaptive learning and extraction of the degree features of the fault and fault recognition; employing the advantage that the sparse automatic coding can automatically learn the internal features of data, and adding noise coding to be integrated in the sparse automatic coding for improving the robustness of feature learning; carrying out the layer-by-layer no-supervision adaptive learning and supervision fine tuning of the original input complex data through multilayer sparse automatic coding, completing the automatic extraction and expression of the degree features of the mechanical fault and achieving the intelligent diagnosis of the degree of the fault. The method is used for the diagnosis of the degree of faults of rolling bearings under different work conditions, and obtains a good effect of feature extraction and diagnosis.
Owner:CHONGQING JIAOTONG UNIVERSITY
Depth significance-based remote sensing image rapid retrieval method
ActiveCN106909924AAchieving the benefits of incremental learningSave storage spaceCharacter and pattern recognitionNeural architecturesResearch ObjectNetwork structure
A depth significance-based remote sensing image rapid retrieval method is disclosed and belongs to the field of computer vision. The method disclosed in the invention specifically relates to technologies such as in-depth learning, significance object detection, image retrieval and the like. According the method, remote sensing images are research objects, and in-depth learning technologies are used for researching a remote sensing image rapid retrieval method. A full convolution neural network is adopted for constructing a multitask significance object detection model which is used for doing significance detection tasks and semantic segmentation tasks at the same time, and depth significance characteristics of the remote sensing images are learnt in network pre-training processes. A depth network structure is improved, a Hash layer fine tuning network is added, and binary system Hash codes of the remote sensing images can be obtained via learning. Significance characteristics and the Hash codes are used comprehensively for similarity measurement. The method disclosed in the invention is of high application value for realizing accurate, highly efficient and feasible retrieval of the remote sensing images.
Owner:BEIJING UNIV OF TECH
Who we serve
- R&D Engineer
- R&D Manager
- IP Professional
Why Patsnap Eureka
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com