106results about How to "Comprehensive training" patented technology

The invention discloses a multi-pose lower-limb rehabilitation robot and a use method thereof, and belongs to the field of medical apparatuses. The rehabilitation robot comprises a pedestal, a liftingmechanism and a bed body which are arranged sequentially from bottom to top; a pair of thigh training mechanisms and a pair of leg training mechanisms are symmetrically arranged on the two sides of the bed body correspondingly; the thigh training mechanisms and the leg training mechanisms are used for performing lower-limb rehabilitation training on a patient whose lower limbs have dyskinesia; the patient is subjected to lower-limb rehabilitation training by multiple modes; furthermore, the bed body and a seat cushion are overturned, so that training of three poses of sitting, lying and standing can be realized, the training mode is comprehensive and the training effect is good. By the method adopting the multi-pose lower-limb rehabilitation robot, the training effect is excellent.

The invention discloses a dynamic gesture recognition method and system based on a two-dimensional convolutional network, and the method comprises the steps: collecting an actual dynamic gesture video, and carrying out the framing processing of the video; carrying out frame sampling on the actual image after framing; encoding the actual image after frame sampling to obtain an actual feature vectorof the actual image; fusing the actual feature vectors to obtain an actual feature matrix; and inputting the actual feature matrix into the trained two-dimensional convolutional neural network, and outputting a gesture recognition result. By processing a source video stream into a frame of image, and sending the frame of image into a two-dimensional convolutional network, a classification resultof a gesture action is obtained; wherein the image generated by the video contains the spatial feature information and the time sequence information of the video. According to the method, the calculation complexity of gesture recognition is effectively reduced.

The invention discloses a neural network training method based on Gaussian process prior guidance. The neural network training method is used for improving the training process of a neural network toobtain a better training effect. The neural network training method specifically comprises the following steps: S1, obtaining a data set for neural network training, selecting a representative set formodeling priori knowledge, and defining an algorithm target; S2, performing a training process of first-stage batch iterative learning on the neural network model, and sequentially executing the steps S21 to S24 in each iterative batch; S3, after the current training process is finished, verifying the neural network model by using the verification set to obtain a verification set error rate of the current model; and S4, continuously repeating the steps S2 and S3 to perform a multi-stage training process on the neural network model until the model converges. According to the neural network training method based on Gaussian process prior guidance, the effectiveness of training can be effectively improved in tasks, and the network learning ability and the learning quality are improved, and the neural network training method has a good application value.

A system for simulating electrical systems of a diesel-electric locomotive comprising includes a first electric motor for simulating a locomotive diesel engine and a generator coupled to the electric motor for simulating a locomotive main generator in a main generator operating mode. A second electric motor is driven by an electrical output of the generator in the main generator operating mode and simulates a locomotive traction motor. A third electric motor is coupled to the second electric motor and operates as a generator providing a load on the second electric motor during simulated locomotive motoring operations.

The invention discloses a word weight predicting model generating method for predicting the word weights in position titles, a position recommending method and a computer device. The method includes the steps of obtaining a pre-collected position title training set, wherein the position title training set comprises a plurality of position titles, each position title is related to a corresponding word separating sequence, and each word separating sequence comprises one or more words obtained after word separating treatment of the corresponding position title; obtaining word features of all thewords in the related word separating sequence for each position title; calculating the marking weights of all the words in the word separating sequence related to each position title; establishing a logic regression model; training the logic regression model with the word features of all the words in the word separating sequence related to each position tile as the input and the corresponding marking weights as the output to generate a word weight predicting model, and indicating the predicted weights of all the words in each input position title through the output of the word weight predicting model.

The invention discloses a multifunctional leg rehabilitation trainer which comprises a bed frame and a bed cushion arranged on the bed frame. A seat board is arranged at one end of the bed cushion onthe bed frame, two symmetric seat cushions are rotatably connected onto the seat board, leg recovery mechanisms are mounted on the seat cushions, and a load increasing module is arranged at a foot recovery module on the bed frame. The leg rehabilitation trainer is simple in structure and can perform knee joint motion training, hip joint motion training, rotation training taking a hip joint as an axis, thigh outside and partial thigh inside muscle group training and rotation training taking a thing as an axis on a patient by the aid of the leg recovery mechanisms, and diversified training of legs can be realized, so that the patient can rapidly achieve leg rehabilitation. In addition, load can be increased by the load increasing module according to practical training requirements to performtraining.

The invention discloses an upper limb rehabilitation training apparatus, which comprises supporting legs, a seat cushion and a backrest, wherein the seat cushion is arranged at the upper ends of the supporting legs; an inserting groove that the cross section is of a non-circular structure is arranged in the back side of the seat cushion in a vertical direction; a first gear rack, which is slidingly connected to the interior of the inserting groove, is arranged at the lower end of the backrest; a first driving motor is arranged on the lower surface of the seat cushion by virtue of a supportingplate; a first gear, which is meshed with the first gear rack, is arranged on an output shaft of the first driving motor; and arm recovery mechanisms are arranged at the left and right sides of the upper end of the backrest. The upper limb rehabilitation training apparatus is simple in structure, and moreover, diverse training of upper limbs can be achieved, so that the upper limb rehabilitation of a patient can be completed rapidly.

The invention discloses a teaching model for abdominal examination skill training during pregnancy. The teaching model comprises a pregnancy abdomen phantom, wherein the pregnancy abdomen phantom can be fixed to the abdomen of a whole female body mode, a uterus phantom and a pelvis inlet phantom are arranged in the pregnancy abdomen phantom, a fetus phantom is arranged in the uterus phantom, and liquid is filled between the fetus phantom and the inner wall of the uterus phantom. A first microphone is arranged on the fetus phantom and used for playing fetal heart sound and umbilical cord sound, a second microphone is arranged on the uterus phantom and used for playing uterine souffle, placental souffle and aorta abdominalis souffle, and the first microphone and the second microphone are in signal connection with a player. A vibration device for simulating fetal movement is arranged on the fetus phantom and is in signal connection with the controller. The teaching model for abdominal examination skill training during pregnancy has the advantages of being simple in structure, good in simulation performance and good in teaching effect.

The invention discloses a remote aid method and a remote aid system. A recourse party and a support party send out a communication connection each to a service party respectively and the connections are successful only after passing a safety certification party. The service party adds the recourse party and the support party to a recourse party queue and a support party queue; the recourse party sends out a technical aid request to the service party via network when a technical aid is needed; the service party dispatches a support part which can provided the technical support from the support party queue, transfers the communication address to the recourse party requesting for aid, and provides a safety certification command to the support party and the recourse party; after the request for remote control connection sent out by the recourse party to the service part passes the safety certification, the support party can provide remote technical support service and the support party and the recourse party can save all the information during the aid locally; after remote aid is completed, the recourse party or the support party initially ends the communication and releases the connection. The invention allows the technical aid to be convenient and rapid without being constricted by working locations.

The invention discloses a leg rehabilitation training device, comprising a bed frame and a mattress arranged on the bed frame. One end of the mattress on the bed frame is provided with a seat plate. And two symmetrically arranged seat mattresses are rotatably connected on the seat plate. A leg recovery mechanism is arrange on that seat cushion; The leg rehabilitation training device not only has simple structure, Moreover, the leg recovery mechanism can realize the knee joint movement training, the hip joint movement training, the hip joint rotation training, the thigh lateral and part of thethigh medial muscle group training, the thigh rotation training. And then can realize the leg diversification training, so that the patient can quickly complete the leg rehabilitation.

The invention relates to the technical field of image processing, in particular to an image target detection model training method and device based on consistency negative samples. In order to solve the problem that in the prior art, a fixed negative sample is adopted, so that a model difficultly learns features with discrimination force, the invention provides an image target detection model training method, which comprises the steps of obtaining an initial image sample set based on the overlapping degree of a real frame in a to-be-identified image and a preset initial anchor frame; accordingto the initial image sample set, through a preset image target detection model, obtaining the prediction anchor frame corresponding to the initial anchor frame, and based on the overlapping degree ofthe real frame and the prediction anchor frame, obtaining an updated image sample set; and training the image target detection model through the updated image sample set. By using the method and thedevice of the invention, the image target detection model can be trained by using more comprehensive information.

The invention discloses a football training device which comprises a bracket, a first ball returning mechanism and a second ball returning mechanism, wherein the first ball returning mechanism is used for enabling a football to return at a low position; the second ball returning mechanism is used for enabling the football to return at a high position; both the first ball returning mechanism and the second ball returning mechanism are arranged on the bracket; the first ball returning mechanism is provided with an elastic ball returning part; the second ball returning mechanism is provided with an arc ball returning part. When a football player is subjected to football training, by the elastic ball returning part and the arc ball returning part, not only can automatic ball returning of a football be implemented and time and physical loss caused by the case that the player picks up the football by the self is saved so as to greatly improve training efficiency, but also training on conventional contents, such as passing, shooting, heading and the like, can be comprehensively carried out, a ball sense and a ball touch level of the player can be rapidly improved, and a training effect on the player is comprehensively promoted.

The invention discloses a badminton ball pushing training monitoring and evaluating system which comprises a badminton ball training field. A ball height measuring device is arranged at a ball net onthe badminton ball training field. A training area for the ball pushing training of athletes is arranged on one side of the ball net. A device area for ball shooting and ball launching position detecting is arranged on the other side of the ball net. The device area comprises a ball falling sensing device located on the rear field and a ball shooting device located outside the rear side of the rear field. A displayer for displaying ball striking scores, a central processor and a power supply device are arranged on the rear side of the ball shooting device. The athletes start the system by pressing a system starting button on a bat. The badminton ball launching position, the flight height in the air and the flight speed can be automatically measured, the ball striking practice is scored according to the three indexes, the athletes can know the ball striking effect of themselves, and the training aim is achieved.

The invention relates to a basketball training system. The basketball training system comprises a ball placing device, a plurality of obstacle handrails, a ball passing training device, a basket ballentering detection assembly and a data acquisition and analysis center. The ball placing device starts training timing and sends a termination moment to the data acquisition and analysis center. The obstacle handrails are arranged along a straight line. The ball passing training device comprises an electronic fence and a ball passing detection assembly, the electronic fence has a function of limiting the ball passing distance of a player, and the ball passing detection assembly D detects whether the player hits a target or not through ball passing. The basket ball entering detection assembly detects whether the player throws a basketball into a basket or not. The data acquisition and analysis module acquires data in real time to obtain the time consumed in the whole training or testing process of the player and the time consumed in each stage. According to the invention, the training and comprehensive test of the overall basic ability of players including basketball dribbling, passingand shooting are realized, and the problems of unscientific training mode and non-objective assessment mechanism in the basketball field at present are solved.

The invention discloses an upper limb rehabilitation training device with a shoulder massage function, comprising support legs, a seat cushion and a backrest, wherein, the seat cushion is installed onthe upper end of the support legs; the rear side of the seat cushion is provided with a socket groove with non-circular cross section along the vertical direction; the lower end of the backrest is provided with a first frame slidably connected in the socket groove; a first driving motor is mounted on the lower surface of the seat cushion through a support plate, and a first gear meshing with thefirst frame is mounted on an output shaft of the first driving motor. The left and right sides of the upper end of the backrest are provided with an arm restoring mechanism; and the left and right sides on the back side are provided with shoulder massage mechanisms. The upper limb rehabilitation training device with a shoulder massage function not only has simple structure, but also can realize diversified training of the upper limb, so that patients can quickly complete the rehabilitation of the upper limb.

The invention discloses a fight training device which comprises a beating column, a driving motor is arranged in the beating column and connected with a rotating block, a swing piece is arranged on the rotating block, and a hanging part is arranged on the swing piece. Through mutual cooperation of a driving motor, a rotating block, a swinging piece, a connecting piece, a hanging ball and an inserting rod, initiative striking and 360-degree rotary boxing swinging can be carried out, then boxing of a person is simulated, corresponding dodging and counterattack training is carried out on trainees, through control of a built-in intelligent program, movement modes of different frequencies are achieved, and the training effect is good. The swinging piece and the rotating block are matched to simulate the arms of a person to attack in the upper direction, the middle direction and the lower direction, the hanging ball is driven by the flexibly-connected connecting piece to throw under the action of centrifugal force, and the fist throwing of the person is simulated, so that compared with a traditional trainer which is single in function, the trainer can actively punch and attack, and the training efficiency is improved. In addition, the training equipment is also suitable for training all standing combat skills such as boxing, taekwondo and karate.

The invention discloses a basketball shooting training device and a wearing method of the basketball shooting training device. A magnetic stick is installed on the back face of a first phalanx of a middle finger of a shooting glove, reed pipes are installed at the periphery of the magnetic stick, normally-opened type reed pipes or normally-closed type reed pipes or simultaneously-installed the normally-opened type reed pipes and the normally-closed type reed pipes are arranged on the back face of a first phalanx of an index finger and the back face of a first phalanx of a ring finger, the back face of a metacarpal bone of the middle finger and the back face of a second phalanx of the middle finger, and when the reed pipes are all communicated and the communication continuous time reaches the set time t1 second, a first single chip microcomputer continuously outputs a low level for the time t2 second, and an output device obtains power to conduct working so as to remind a trainer that the hand shape is correct. When the communication continuous time of the reed pipes does not reach the set time t1 second or any one or more than one reed pipe is not communicated, the output device does not obtain the power so as to stop working. The basketball shooting training device not only can independently train the shooting hand shape, but also can be combined with a palm horizontal meter, a front arm vertical meter and an upper arm horizontal meter to complete three 90 degrees and stimulate lower limb strength training, training is more complete, and the basketball shooting training device is more convenient to wear.

The invention provides a fine action training system and method. The system includes a display device; the display device is respectively connected with a peripheral operation device and a data processor, a user operates the peripheral operation device according to the information displayed by the display device, and the operation information is uploaded to the data processor; the data processor compares the operation information with a standard trajectory, conducts scoring by a scoring module, and displays a scoring result on the display device. The system expresses a training result of userfine action visually through specific and scientific scoring, and can provide a corresponding correcting scheme according to the progress situation; meanwhile, the system can also depict the trajectory of ability improvement to ensure the comprehensive, objective and accurate fine motion training of children.

The invention discloses a rehabilitation training robot and a rehabilitation training device having the same. The rehabilitation robot comprises a robot body and a robot arm. The robot arm is connected movably with the robot body, and used for the rehabilitation training. A foot fastener is connected with the robot arm, and used for fastening the feet. The foot fastener is provided with a massaging device. According to the rehabilitation training robot, the massaging device is arranged on the foot fastener to add a massage mode in the lower limbs training process, and make the training at one same time point more comprehensive. The patient not only can train the joints and muscles of the lower limbs in a large scope and various modes, but also can do the foot massage to benefit the blood circulation of the lower limbs, and the rehabilitation of the patient.

The invention discloses an image target detection method and device, terminal equipment and a computer readable storage medium, and the method comprises the steps: marking a label of a to-be-detectedtarget in a target sample image containing the to-be-detected target; constructing each training sample image according to the target sample image marked with the label and the scene sample image of each scene field, each scene sample image not containing the to-be-detected target; inputting each training sample image into a preset target detection model for model training, and detecting an imagetarget in the to-be-detected image by using the trained preset target detection model. According to the invention, there is no need to spend a lot of manpower resources to collect image data in different scene fields and mark image targets; training sample images for model training of the target detection model are enriched, false detection and missing detection of the target detection model during subsequent image target detection are avoided, and the image target detection efficiency is improved.

The invention discloses a top drive simulating system. The top drive simulating system comprises a top drive simulation operation system, a PLC (programmable logic controller) controller and a control; the control unit comprises a main control computer, a parameter machine and a graphic processor; the top drive simulation operation system is connected with the PLC controller, and the PLC controller is connected with the parameter machine through a serial port; the parameter machine and the graphic processor are respectively connected with the main control computer through a communication network. The top drive simulation system can simulate all operation of a real top drive, control and display elements on the operation table are consistent with the actual control table, thus all parts of the top drive device can be controlled, related parameter can be displayed in real time according to actual operation condition, so as to reach the effect as the on-site operation; besides, the top drive simulating system can perform comprehensive training and examination on top drive constructors, overcome shortcomings of big training difficulty, bad safety, and difficult examination of the traditional training method, and provide a set of scientific, standard and high-efficient new pattern for training and examination.

The invention relates to a smart simulation method for an indoor fire simulation training scene. The method includes steps of establishing a flue-gas temperature model before indoor fire flashover; obtaining a curve of flue-gas temperature varying along with time according to the flue-gas temperature model and predicating flashover time; establishing combustion combinations with different fire extinguishing difficulties; applying the flashover time obtained through prediction and the combustion combinations with different fire extinguishing difficulties to a fire scene environment. According to the invention, the flue-gas temperature model can be used for predicating flashover time of a plurality of indoor scenes such as offices, hospitals, theaters and the like and flashover time in two conditions including different positions in a room and different combustibles of an initial fire source can be predicated at the same time, so that multiple fire scene environments can be provided forfire control simulation training and training diversity is achieved.