59 results about "Circuit training" patented technology

A cordless surgical device includes a modular battery, a radio-frequency signal generating assembly, a surgical handle and an interchangeable circuit casing. The RF signal generating assembly includes RF-signal-generating circuitry, a voltage-control circuit and an external output. The voltage circuit is configured to control an output of the RF-signal-generating circuitry. The handle is configured to support a bipolar end effector having jaws with bipolar contacts and a cutting blade disposed between the jaws. The handle includes leads operable to receive RF signals from the external output and defines an aseptically sealable battery-holding compartment configured to hold the battery. The circuit casing is configured to house the RF signal generating assembly and includes a securing connection adapted to couple the external output of the RF signal generating assembly to the leads. The external output is operable to impart RF signals to the handle when the circuit casing connects to the handle.

Techniques are provided for determining a circuit training path in a smart gym for an exerciser to perform an exercise program in order to minimize user contention for exercise equipment. The training path is defined by the exerciser visiting a number of exercise machines in the order specified by a system. The system determines a next exercise machine for the exerciser to visit which addresses the personal exercise program. In making this determination, the system communicates with all station agent nodes in the system to determine the set of station agent nodes that have an associated exercise machine which operate to address the personal exercise program. Out of that set of station agent nodes, the system reserves the next exercise machine associated with one of the set of station agent nodes that has a state value reflecting the least waiting time for the exerciser.

A system and method for efficient automation of a fitness circuit is disclosed. The system includes a number of networked client computers positioned at workout stations. The client computers are coordinated by a primary client computer which serves as a relay between the client computers and a centralized circuit control system where circuit scheduling and workout details are defined. The method includes enabling an exerciser to scan an issued identification card at each workout station. A client computer positioned at the workout station retrieves and displays workout information specific to the user and further instructs user when to progress to the next circuit station.

The invention relates to a training optimization pornographic picture or video detection method based on a convolutional neural network. A final detection model is obtained through circuit training, classification detection accuracy can be improved to a largest extent, and deficiencies that a perfect complexion model and high detection accuracy are difficult to establish due to ambient light diversity and race diversity and the detection accuracy is always low due to external factors, such as human body gesture diversity, sheltering and the like can be overcome. When the method disclosed by the invention is used for picture data or video data to detect pictures or a video image sequence, only a classification result shows that the pictures or video image sequence is pornographic, a picture set or videos are defined as pornographic data, and a deficiency that a universal rule is difficult to establish for different pornographic videos is overcome.

The invention relates to an image classification method based on direct push semi-supervised depth learning, which comprises the following steps: preparing a semi-supervised image data set, and dividing the training data into a training data set and a verification data set, wherein the training data set has a part of data labeled and the other part unlabeled; verifying the labeled data; on the labeled training dataset, training the general depth neural network image classification model. When the trained model achieves the expected precision on the verification dataset, the network model parameters are saved. A direct push semi-supervised depth convolution neural network model based on the Min-Max principle is built and both labeled and unlabeled data in the training data set are used totrain the model circularly; when the number of cycles reaches the maximum number of cycles, the parameters of the network model are saved. The trained model is used to calculate the recognition accuracy of the label or test data set of the test image. The TSSDL algorithm provided by the invention has good portability.

The invention discloses a multi-person posture estimation method based on adversarial learning. The multi-person posture estimation method comprises the following steps: employing a public data set with a multi-person key point coordinate label as a training set, and carrying out the edge information enhancement preprocessing of a training set image; preprocessing the key point coordinate tags inthe training set, and making a corresponding key point hotspot map and an overall skeleton hotspot map; constructing a double-branch key point feature extraction sub-network; constructing an A-HPose network generator part; constructing an A-HPose network discriminator part; performing relay supervision loop training on the A-HPose network by using the training set to obtain network model parameters; and carrying out post-processing on the network output hotspot map, carrying out search classification on key points in the key point hotspot map according to the skeleton hotspot map to obtain a key point position of each person in the plurality of persons, and estimating the postures of the plurality of persons. The multi-person posture estimation method has the beneficial effect of quickly and accurately detecting human body key point features.

The invention discloses an ultra-short-period photovoltaic prediction method. The method comprises the following steps: selecting training data x; performing normalization processing on the training data; performing data exception handling on the training data; performing data functional transformation; performing significance analysis; training a generalized regression neural network model; and predicating the generalized regression neural network model. According to the ultra-short-period photovoltaic prediction method, a generalized regression neural network modeling theory and method is adopted; partial approximation is further accurate by adding a primary function in a hidden layer, and global optimum is achieved; significance extraction and improvement is carried out specific to the model input information; the correlation of historical data is enhanced through the functional transformation, and the historical data, used as the input signal, enters the generalized regression neural network prediction model, so that the prediction efficiency is effectively improved; in addition, after a training sample is chosen, the generalized regression neural network structure and the weight are determined automatically by only requiring to adjust smoothing parameters, so that the computational process for circuit training is avoided, and the global approximation study and prediction capability is realized more rapidly.

A hand driving cycle exerciser includes a base having a front wheel device and a rear fork to support a rear wheel with an axle. Two rotary members are rotatably attached onto the rear wheel with unidirectional bearings, two followers are rotatably engaged onto the axle and coupled to the rotary members, for unidirectionally rotating the rear wheel via the rotary members and the unidirectional bearings, two handles are rotatably attached to the base and connected to the followers, to unidirectionally rotate the rear wheel via the followers and the rotary members and the unidirectional bearings with the handles.

The invention discloses an image recognition method, a training method and a device of a convolution neural network model. The invention relates to the technical field of water body recognition and can solve the problems of low accuracy and low efficiency of water body recognition in remote sensing images in the prior art. The image recognition method comprises the following steps: preprocessing the remote sensing image to be measured to obtain the multi-feature fusion data; The multi-feature fusion data is input into the trained convolution neural network model for recognition to obtain the water body image. The training method of convolution neural network model includes: acquiring multi-feature fusion data of training remote sensing image; Multi-feature fusion data is used to train theinitialized convolutional neural network model to obtain the training output categories. Adjusting the model parameters according to the errors of the training output category and the mark category; The model of convolution neural network is obtained by training and adjusting the parameters of convolution neural network. The invention is widely applicable to water body identification scene of remote sensing image.

The invention discloses a multi-domain image conversion method and system based on a generative adversarial network. The multi-domain image conversion method comprises the steps of inputting an original image x and an original image y of a specified X mode and a specified Y mode; performing encoding and decompression on the original image x and the original image y in the reconstruction training part to obtain original image characteristics, reconstructed images and reconstructed characteristics respectively, and performing modal discriminant adversarial learning on the characteristics and the images; enabling the loop training part to generate a reconstructed graph, reconstructed graph features and a loop reconstructed graph based on an encoder of an original graph feature exchange modeof a preamble, performing modal discrimination confrontation learning of the features and the graph again, and finally outputting the loop reconstructed graph. A semi-supervised learning method is adopted, existing label data can be used, label-free data can also be used, multi-directional multi-domain image conversion can be achieved without being limited to one-way domain conversion or two-way two-domain conversion, the number of domains is not limited, and the problems of image style migration, medical image multi-mode conversion and the like can be solved.

The invention relates to the technical field of medical instruments, in particular to a limb coordination rehabilitation training device for a neurology department. The limb coordination rehabilitation training device comprises a base, a seat, driving devices and a leg training device. The driving devices are positioned on two sides of the base which is positioned at one end above the base, and the leg training device is positioned at the other end above the base. The driving device comprises a left driving part and a right driving part, the leg training device comprises a left training part and a right training part, the left driving part is used for driving the right training part, and the right driving part is used for driving the left training part. An ultrasonic massaging part is arranged on the seat. By right leg training under left hand drive and left leg training under right hand drive, accordance with human nervous system tissues is realized, and coordination training of upperand lower limbs of patients is benefited; high training pertinence and great recovery effects are achieved; by the ultrasonic massaging part for ultrasonic massaging of thighs and cervical vertebra of the patients, blood circulation is promoted, metabolism rate is increased, and recovery of the patients is benefited.

The invention discloses a two-way colorization method for animation images based on a U-shaped periodic consistent confrontation network, relating to the Image processing field, Data collection, pixels of the animation illustration image are unified; A training data set and a test data set are constructed, finally, the capacities of a generator G, a generator F, a discriminating network DX and a discriminating network DY are improved by adopting a cyclic training method with consistent U-shaped periods, a function mapping relation between an image and a color image is found, and bidirectionalconversion of a black-white sketch and a full-color image is realized. According to the method, the features do not need to be extracted manually, a training set does not need to be marked, the workload of a cartoon creator is remarkably reduced, the image colorization and blackening processing efficiency is improved, and great help is provided for the cartoon creator.

The invention discloses a method and a system for low-altitude ground vehicle detection and motion analysis, in which a positive sample and a negative sample are captured in a video frame in advance. The method comprises the follow steps of: performing circuit training on a training sample according to characteristic blocks divided by the training sample to obtain a weak classifier of each regionin each cell gradient direction of the training sample; combining the weak classifiers to obtain a strong classifier corresponding to each cell; taking output values of the all the strong classifiersas feature vectors, training to obtain a support vector machine classifier, detecting the vehicle by using the support vector machine classifier, and marking an area as a vehicle area when the vehicle is detected in the area; calculating color space similarity between the vehicle area and an image area scanned by a scanning window in neighboring image frames; and comparing the space color similarity to obtain the motion trail of the same vehicle, and performing motion analysis. The method can improve accuracy rate of low-altitude ground vehicle detection, and also can improve accuracy of vehicle motion analysis.

A computer-implemented system for improving a circuit training exercise routine at an exercise facility having at least two pieces of exercise equipment. The system using a smart device, and having steps including: initiating a program on the smart device upon use of a first piece of exercise equipment; displaying exercise instructions to the user on the smart device; initiating a timer on the smart device with a pre-selected amount of time in which the user must complete the exercise instructions on the first piece of equipment; recording if the user is successful in accomplishing the first set of exercise instructions; and prompting the user to move to a second piece of exercise equipment. The steps may also include displaying additional exercise instructions. The system includes monitoring the user's health functions and rewarding the user for successful accomplishment of the first and additional sets of exercise instructions.

The invention relates to an infrared video colorization method based on two-channel-circulation generative adversarial network. The method comprises the following steps of collecting night-vision infrared video data and color video data; carrying out frame extraction processing on a video data set and then putting into a database, and building a training set and a test set; constructing a two-channel-circulation generative adversarial deep learning network, and through a circulation training method, improving the abilities of generating a network G, generating a network F, discriminating a network DX, and discriminating a network DY; and acquiring an infrared image in an infrared video in real time, inputting the colorization results of the infrared image and the previous frame of infrared image into the trained and generated network G, and using the generated network G to carry out colorization processing on an infrared image sequence. In the invention, the colorization effect of a night-vision infrared video is enhanced, the observability of a colorization video is increased and manual intervention is not needed.

The invention provides a convolutional neural network model based data processing method and device. The method includes steps: cyclically training parameter data of a convolutional layer and / or a full connection layer in a convolutional neural network model to acquire discrete data in a preset format; storing the discrete data in the preset format by preset bits. According to the technical scheme, the parameter data are converted into the discrete data which are then stored by the preset bits, and consequently compression storage of the model is realized while accuracy loss of the model after conversion is avoided; due to adoption of the discrete data in the preset format, operation efficiency is greatly improved.

A convertible cycle exerciser. The exerciser, moveable between a first position and at least one second position, comprises a base and a supporting member extending upwards therefrom with a main exerciser body pivotally connected to the supporting member by at least one pivot point. In the first position, force is applied to crank arms via a user's feet in a conventional cycling mode, and in the at least one second position, applied to the crank arms with a user standing and operating the exerciser via hands and arms. An adjustable resistance mechanism acting on a first rotating member to vary the load requirement is controlled by at least one resistance control conveniently accessed in each of the first and at least one second positions. Movement of the exerciser between the first and at least one second positions is facilitated by operation of a simple release mechanism. Compared with conventional convertible systems, the cycle exerciser presents a considerably simpler mechanism, reducing complexity of the system and thereby cost, weight, and maintenance requirements, providing an additional upper body exercise component with simplified and convenient engagement and resistance adjustment access in all positions of use.

A magnetic circuit for a speaker comprises: a bottom yoke including a center pole; a ring magnet; a top plate; and a short-circuiting ring shaped like a hollow cylinder. The short-circuiting ring is disposed close and parallel to a voice coil and attached to the inner portion of the bottom of the top plate.

The invention discloses a method for training multiclass Boosting categorizers. The method is characterized in that a class weight corresponding to the class attributed to a training sample is distributed for a sample weight of each training sample in the training data after each circulation and before starting the next circulation in the training process, that is, each training weight of the training sample in the training circulation comprises the sample weight and the class weight. The class weight corresponding to each class obtains a strong categorizer according to the circulation training and the latest circulation training towards the performance and the dynamic change of the class, so that the training weight of the training sample of the class with poor performance in the next circulation is enlarged, the training weight of the class with good performance in the next circulation is reduced, and the performance of each class reaches the target threshold of the performance as far as possible in the same circulation to achieve the training. Therefore, the invention eventually ensures that the quantity of the weak categorizer required by the class with the worst performance is reduced; and meanwhile, the quantity of the weak categorizer required for categorizing different classes is basically the same.

The invention discloses a method and a device for identifying sketch face and belongs to the field of face identification. The method comprises the following steps: obtaining face training samples, filtering, carrying out LBP (Local Binary Patterns) processing and blocking, then grouping the face training samples into N groups, and setting weight for each group; training N groups of LBP images; calculating the weighting error rate of each sub-block according to a matching degree, between a sketch image and a visible image, of sub-blocks in the same position and selecting the sub-block with the minimum weighting error rate and then adjusting the weight of each group; circularly training till the identification rate reaches an assigned value and recording the information, obtained by training for each time, of the sub-block with the minimum weighting error rate; calculating weighting LBP distance between the to-be-identified sketch image and each test sample according to the information of the sub-block when the to-be-identified sketch image is input, and taking the test sample with the minimum weighting LBP distance as an identification result. The device comprises an initial processing module, a training module and an identification module. By adopting the method and the device for identifying the sketch face, the sketch face identification complexity is reduced; the identification efficiency and accuracy are improved.

The invention relates to a method for fixing the connecting sequence and character threshold value of cascade classifiers with different characters, wherein said cascade classifiers are used to extract selected connecting components from the candidate connecting components decomposed from image; said method comprises: first decomposing at least one image to obtain several connecting components as present array; then parallel feeding present sample into cascade classifier, to circulated train, to fix the connecting sequence and character threshold value of cascade classifiers. The invention also provides a method for obtaining selected image from one image, which uses the cascade classifiers via aforementioned method, to quickly remove non-selected connecting components, to spend more cost on selected connecting components, to improve the image processing speed and improve the obtaining accuracy.

The invention relates to a method for detecting the appearance of an object and a depth neural network model. The method for detecting the appearance of an object comprises the following steps: building a depth neural network model; deep neural network model has the ability to distinguish the appearance through the cycle of training and learning; according to the large number of cyclic training and learning, the depth neural network model gradually converges to obtain the optimal weight of each eigenvalue; convolution operation is carried out on the appearance picture of the object to be detected, and the classification result is obtained according to the set probability interval value. The depth neural network model comprises a training module, an evaluation module and a prediction module. The invention solves many problems existing in manually formulating judgment rules of object appearance classification, and overcomes the problems of low manual operation efficiency or low accuracyof traditional automatic judgment. At the same time, because of its sustainable self-iterative upgrading, its recognition efficiency will be improved in theory. The module of the invention is simple,the hardware cost is low, and the application range is wide.

A frame synchronization method of orthogonal frequency compound system. The timing list is defined as square of two variables adding ratio, the numerator variable is the module value of the receiving signal and compound signal, the denominator is the instant power of training symbol. This frame synchronization technology is suitable for all orthogonal frequency system allowing adding circular training symbol at the front end of the data frame. It includes the following processes: 1) choose the delay of the receiving signal and the length of relative calculation; 2) calculate the energy of the relative length training symbol; 3) according Monte Carlo result or theoretical results set the expected value of the timing order at the right time; 4) according the expected value of the timing order at the wrong time and the artificial valve value calculate the valve value including test and judgment; 5)calculate the timing list of the receiving data and judge the startup time according to the valve value.

The invention provides a channel estimation method based on a circuit training sequence. According to the method, output matched with a filter (related device) is subjected to interception and zero fill by taking a related peak as a center at a receiving end, so that a time domain denoising effect is achieved; and the sequence subjected to interception and zero fill is compensated in a frequency domain, so that energy loss caused by zero fill is reduced.

The present disclosure relates to a stationary exercise apparatus for indoor cycle training (1), i.e. a stationary exercise bicycle, preferably provided with a magnetic resistance unit (15). One embodiment relates to a stationary exercise bike comprising a flywheel (10) defining a radial gap (25) in the periphery wherein at least the periphery of the flywheel (10) has ferromagnetic properties, and wherein a magnetic resistance unit (15) is configured to controllably insert one, two or more magnets into said radial gap.

A toy exercise facility for children to pretend play that their dolls are exercising at a circuit training facility is discloses. The inventive toy exercise facility includes a convertible carrying case with a top section, a bottom section, and a plurality of side sections. The carrying case may be opened and laid flat. A plurality of toy replica circuit training stations are attached to the carrying case sections such that when the carrying case is completely open, the child may pretend play that her doll is exercising at the several circuit stations. The toy exercise facility also has music playing to emulate the playing of high energy background exercise music at a gymnasium circuit training facility. The carrying case has interconnected sections to allow the case to be configured into a closed carrying case for ease of transport and storage.

The invention relates to a training method of a vehicle logo classification model, a vehicle logo recognition method, device and apparatus. The vehicle logo classification model training method comprises the steps of obtaining a plurality of training sample images, and fusing any two training sample images in the plurality of training sample images according to a preset fusion proportion to obtaina fused sample image; inputting the fused sample image into a vehicle logo classification model for classification processing to obtain a vehicle logo classification result; respectively calculatinga vehicle logo classification result and a first classification loss and a second classification loss of category labels of any two training sample images, and fusing the first classification loss andthe second classification loss according to a fusion proportion to obtain a fusion loss; and adjusting model parameters in the vehicle logo classification model according to the fusion loss, and circularly training until the vehicle logo classification model is converged. Discrete samples can be continuous, the smoothness in neighborhoods is improved, and the problem of overfitting is solved; andmeanwhile, the model training efficiency is improved.

A method for generating a cyclic training sequence comprises the following steps: generating randomly the first information block with the length N, copying the last L samples of the first information block to the beginning of the block as the first CP, generating the second information block with the length N by using the first information block of the training symbol in a circle manner in which the last d samples of the first information block are shifted to the position before the rest (N−d) samples and the N samples thus rearranged forms the second information block, copying the last L samples of the second information block to the beginning of the block as the second CP to form the integrated cyclic training sequence.

An exercise apparatus that provides easily and quickly adjustable resistive forces during a wide variety of fitness-related activities. The apparatus comprises a housing having an internal space wherein multiple resilient members, such as elastic cords or stretchable bands, are fully contained while in a rest state; an attachment point where each resilient member is securely affixed to the housing; and an aperture located substantially opposite the attachment point, through which resilient members are accessible, and to which a handle assembly is removably attached. A user may selectably attach the handle assembly to one or several resilient members in order to select the desired resistive force. Because resilient members are fully contained within the housing when in a rest state, the resilient members exert a resistive force immediately upon being extended from the aperture in the housing.