177results about How to "No loss of accuracy" patented technology

An integrated disposable cartridge for dialysis or ultrafiltration treatment of blood is disclosed that includes integral miniature pressure sensors. Sensors are embedded in the tubing of the cartridge to measure pressure of blood or other fluids. Cartridge elements form a continuous smooth bore passage for blood that reduces probability of clotting.

The invention discloses a method for automatically identifying a breast tumor area based on an ultrasound image. The method comprises the following steps of acquiring the ultrasound image of the breast, and preprocessing the ultrasound image; segmenting the ultrasound image subjected to preprocessing through an image segmentation method to obtain a plurality of segmented subareas; extracting a grey level histogram, texture features, gradient features and morphological features of the ultrasound image, and combining the grey level histogram, the texture features, the gradient features and the morphological features of the ultrasound image with two-dimensional position information to obtain high-dimensionality feature vectors; selecting the most effective feature subset of the high-dimensionality feature vectors through feature ordering based on biclustering and a selection method; performing learning classification on the selected most effective feature subset through a classifier, and then automatically identifying the breast tumor area. By means of the method, the breast tumor area can be identified automatically from segment results of the breast tumor ultrasound image, therefore, automation performance of computer-aided diagnosis is improved, manual operation of clinical doctors is reduced, and subjective influence of clinical doctors is reduced.

The invention discloses a two-dimensional recursive network-based recognition method of Chinese text in natural scene images. Firstly, a training sample set is acquired, and a neural network formed bysequentially connecting a deep convolutional network, a two-dimensional recursive network used for encoding, a two-dimensional recursive network used for decoding and a CTC model is trained; test samples are input into the trained deep convolutional network, and feature maps of the test samples are acquired; the feature maps of the test samples are input into the trained two-dimensional recursivenetwork, which is used for encoding, to obtain encoding feature maps of the test samples; the encoding feature maps of the test samples are input into the trained two-dimensional recursive network, which is used for decoding, to obtain a probability result of each commonly used Chinese character in each image of the test samples; and clustering searching processing is carried out, and finally, the overall Chinese text in the test samples is recognized. According to the method of the invention, space/time information and context information of the text images are fully utilized, the text imagepre-segmentation problem can be avoided, and recognition accuracy is improved.

The invention relates to the internet safety technology, and is designed to provide an intelligent detection method for Java script malicious code facing to the webpage. The method comprises three processes of optimization of sample, safety detecting and renewal of optimization. A classical N-gram statistical model and a KNN grader are efficiently fused by means of the efficient mixing of static detecting and dynamic detecting. The dynamic behavioural analysis of the code is realized by setting up N-gram characteristic of a Java Script machine code operation order. The efficiency of the static classification detection is largely improved by the optimization of the sample relayed on the KNN grader. The suggested intelligent detecting method is specially provided with operability, the step of optimization of sample can ensure that the classification effectiveness is not decreased when a training script is enlarged, and the step of the safety detecting can ensure to execute the high-efficient intelligent detecting based in the optimization of sample, and the step of the renewal of optimization can ensure thag the precision of the intelligent detecting is not be declined when new malicious script is increased. The method has the advantages of providing the ability of detecting the new malicious script and the dynamic optimization adjust ability for continuing work.

A detection device of multiple displacement /deflection is prepared as setting projection target at lens end of shooting equipment, sticking film with accurate scale on back of projection target, setting n numbers of leaser emitters separately at each point to be detected and aiming them to projection target for forming light spot used on shooting equipment detection through face plate.

The present invention discloses an image detection method for cable rope surface defects, and aims to detect if cable rope surface images have defects. The method comprises the steps of performing graying processing on a cable rope surface image, carrying out enhancement processing on the grayed image by using an improved local grey-scale contrast method, carrying out a cutting processing on the enhanced image by using an improved maximum-correlation method, and, ultimately, performing defect determination according to the size of cut regions of the cut image. With adoption of the method, intelligent detection and judgment of cable rope surface defects can be realized, the workload and fatigue of manual visual inspection are reduced, the detection efficiency may be raised, and the method is safe and reliable.

The invention discloses a microblog pushing method and device based on user propagation power, aiming at solving the problems in the prior art that the microblog pushing timeliness is reduced and network resources are wasted caused by low accuracy in the propagation power determination. The method comprises the following steps that: a server determines the forwarding relationship and propagation power transfer ratio among all users according to preset initial propagation power values, determines the propagation power of each user, pre-estimates the hot degrees of microblogs to be pushed according to the determined propagation power of each user, and accordingly pushes the microblogs to be pushed. According to the method, the determined propagation power has no subjective errors, and the situation that the accuracy of propagation power determination is reduced due to the existence of zombie fans can be avoided, so that the propagation power of each user can be accurately determined, the accuracy of pre-estimation of the hot degrees of the microblogs to be pushed is improved, and when the microblogs are pushed according to pre-estimated hot degrees, the microblog pushing timeliness can be improved and the network resources cannot be wasted.

The invention discloses a measurement method for the rotation angle of an automobile steering wheel. A main rotation body in the automobile steering wheel drives an auxiliary rotation body to rotate under the rotation driving of a transmission shaft; the transmission ratio of the main rotation body and the auxiliary rotation body is m:n; the main rotation body gains the relative rotation angle Theta1 of the main rotation body by a first rotation angle sensor; the auxiliary rotation body gains the relative rotation angle Thelta2 of the auxiliary rotation body by a second rotation angle sensor, thus gaining practical intercept K; theoretical intercept L is gained by the practical intercept K; the rotation ring number P of the main rotation body is then gained by the theoretical intercept L; the relative rotation angle Thelta1 of the main rotation body and the relative rotation angle Theta2 of the auxiliary rotation body are input into a micro-processor; after processing, the micro-processor outputs an absolute rotation angle Theta in m ring of the main rotation body, namely, the absolute rotation angle Theta in m ring of the main rotation body is gained by the rotation ring number P of the main rotation body and the relative rotation angle Theta1 of the main rotation body. The measurement method of the automobile steering wheel can eliminate the system error which is caused by idle course and realize precisely measuring the absolute rotation angle Theta in m ring in small volume and simple structure.

The present invention provides a robot system and a map updating method. Because of map update, error is generated in previous environment map and an updated environment map. Therefore, when update processing is performed repeatedly, errors are accumulated. A plurality of pairs which are composed of shape information of a preset area in the environment map that is preset and the property of the preset area in the environment map are stored. The environment map is compared with the preset area which is stored in a storage part. The property which corresponds with the preset area is read out. The environment map is updated according to the property in a measuring part.

To determine whether an exhaust gas purification catalyst (42) of an engine (1) has deteriorated, a characteristic of exhaust gas on an upstream side of the catalyst (42) is detected, a characteristic of the exhaust gas on a downstream side of the catalyst (42) is detected, and when the engine (1) operates at a lean air-fuel ratio, a value expressing a difference in the oxygen concentration on the upstream side and downstream side of the catalyst (42) (an average excess air ratio difference) is calculated from the two detected values. When the value expressing the difference in the oxygen concentration is smaller than a deterioration determination threshold calculated in accordance with a quantity of state of the catalyst (42), the catalyst (42) is determined to have deteriorated.

Provided is an apparatus for modeling a three-dimensional object configured to model a three-dimensional object by a fused deposition modeling, including a plurality of material resin supply units configured to supply filaments to be used as a modeling material, respectively, a mixed resin ejection unit configured to eject a mixed resin, which is obtained by mixing the filaments to be supplied from the supply units, and a control unit configured to control amounts of the filaments to be supplied from each of the supply units to the ejection unit. The supply units are configured to supply the filaments of different colors to the ejection unit, respectively, and the control unit is configured to control the amounts of the filaments to be supplied from each of the supply units to the ejection unit, thereby adjusting a color of the mixed resin that is to be ejected by the ejection unit.

The present invention discloses a method for counting mobile devices. The method is performed in a calculation device. The method comprises the steps of receiving a user request, wherein the user request comprises specific region information, specific time information, one or more fence areas, and one or more specific conditions; determining a location code set corresponding to each fence area of the one or more fence areas; according to the specific region information and specific time information in the received user request, extracting multiple pieces of target mobile device information from multiple pieces of stored mobile device information; for all extracted target mobile device information, converting each piece of feature information into multiple bitmaps with location codes as indexes; selecting a desired bitmap according to the location code set and specific conditions; and performing bit operation on the selected bitmap so as to acquire the quantity of mobile devices meeting the user request. The present invention further discloses an apparatus capable of implementing the method, and a calculation device including the apparatus.

The present invention discloses a Chinese text overall recognition method in a natural scene image. The method comprises: acquiring a training sample set, and training a neutral network formed by sequentially connecting a deep convolution network, a multi-layer bidirectional recursive network for encoding, a multi-layer bidirectional recursive network for decoding, and a CTC model; inputting a test sample into the trained deep convolution network to obtain each feature sequence frame of the test sample; inputting each feature sequence frame of the test sample into the trained multi-layer bidirectional recursive network for encoding to obtain each code sequence frame of the test sample; inputting each code sequence frame of the test sample to the trained multi-layer bidirectional recursivenetwork for decoding, and obtaining each common Chinese character in each frame image of the test sample as a probability result of the Chinese character in the frame image; and carrying out cluster search processing to finally recognize the overall Chinese text in the test sample. According to the method disclosed by the present invention, the accuracy of the Chinese text overall recognition in the natural scene image is improved.

Embodiments of the invention provide a target detection method and device, equipment and a storage medium. The method comprises the following steps: subjecting point cloud data to mapping to a two-dimensional space to obtain a first mapped image, carrying out fusion on the first mapped image and a scene image, carrying out target detection in the obtained fused image, and subjecting the obtained detection result to inverse mapping to a three-dimensional space so as to obtain a target detection result. The fused image is two-dimensional data, and target detection is carried out in the two-dimensional data, so a calculation amount is reduced; the fused image comprises data of two aspects, namely the point cloud data and the scene image, and target detection is carried out based on the data of the two aspects, so the accuracy of the detection result is improved; the detection result in the fused image is inversely mapped to the three-dimensional space to obtain the detection result in thethree-dimensional space; and therefore, with the above scheme, under the condition that the accuracy of the detection result is not reduced, the calculation amount is reduced, and the detection result in the three-dimensional space is obtained.

The invention discloses a power line inspection image automatic identification method based on a neural network and a power line inspection image automatic identification system thereof. The method comprises the steps that input images acquired in power line inspection are acquired, and the input images are inputted to a convolutional neural network; 2) a feature vector group is obtained, and the input images apply the feature vectors in the feature vector group in turn so that digital description of the select box of the target type and the target position is obtained; and 3) convolutional operation is performed on the target input image so that the select box of the target object and the target position is obtained, and the object and the position expected to be selected of the highest confidence are also obtained. An end-to-end training model is established so that the target object in the images can be automatically identified, power line inspection image automatic identification can be realized based on the neural network and the efficiency can be enhanced.

A microwave power transducer of two-ends heating type is prepared as using GaAs as substrate, arranging a layer of AIGaAs film on said substrate, setting thermoelectric stack formed by a group of series-connected GaAs/AuGe thermocouple on AIGaAs film, using coplanar waveguide A on GaAs substrate as input transmission line of microwave signal to be measured and connecting coplanar waveguide with two parallel-connected DC heating resistances.

The invention discloses a nested disc cam focusing mechanism, belonging to the field of cam focusing. The nested disc cam focusing mechanism provided by the invention solves the problems that a focusing mechanism of a nested optical system is complicated, the focusing precision is low and the like. The mechanism mainly comprises a first microscope base, a second microscope base, a third microscope base, a fourth microscope base, a fifth microscope base, a sixth microscope base, a slider, a guide rail, disc cams, a driven member, a spring and the like, wherein the first microscope base and the sixth microscope base form a first microscope base group; the second microscope base and the fifth microscope base form a second microscope base group; the third microscope base and the fourth microscope base form a third microscope base group; the microscope base groups move relatively along the direction of an optical axis to form a nested structure; three disc cams control the moving rules of the three groups of nested microscope bases through the driven member respectively and change the positions of the three groups of nested microscope bases so as to realize focusing of the whole system; and moreover, due to the independent microscope base and disc cam structure, compared with a cylindrical cam, the structure is simple, the processing cost is low, and the focusing precision is high.

Cone-beam reconstruction is performed within a practically acceptable time on a computer having one or more microprocessors. The calculations involved in the reconstruction are divided into calculations to be performed on the MMX, ALU and SSE units of each of the microprocessors. For pure floating-point data, it is preferred to use the MMX unit to adjust the data address and map data and to use the SSE unit to perform the backprojection. The data are partitioned by z-line so that the data to be processed in each stage of the backprojection fit within the L1 cache.

The invention discloses a voiceprint recognition poisoning sample generation method based on migration training. The voiceprint recognition poisoning sample generation method comprises the following steps: (1) preprocessing a voice data set; (2) building a voiceprint recognition model; (3) obtaining a feature representation space of the migration training task data set; (4) selecting a target sample and a base sample from the test set; (5) generating a poisoning sample by using an optimization algorithm; and (6) adding the poisoning sample into the original training set for migration training: under the condition of not changing the weight of the original model, only retraining the replaced softmax layer to be suitable for the current classification task, and adding one poisoning sample into the original training set by the training set. In a test phase, a target sample will be misclassified as a tag of a base sample. According to the invention, the attack success rate is greatly improved.

The invention discloses an electronic equipment fault prediction method based on an LSTM enhancement model, and relates to a method for performing memory enhancement on a long-short-term neural network. In order to overcome the problem that the precision is gradually reduced along with the time sequence length when an existing LSTM model is adopted to predict an equipment fault, the method specifically comprises the following steps: step 1, collecting monitoring data in the operation process of electronic equipment, and preprocessing the monitoring data; 2, building an LSTM enhancement model,and inputting the preprocessed monitoring data into the LSTM enhancement model for training to obtain a trained LSTM enhancement model; and step 3, according to the trained LSTM enhancement model andthe monitoring data in a period of time, predicting the fault of the electronic equipment.