455results about How to "Efficient and accurate detection" patented technology

The invention discloses a deep-convolution-neural-network-based CT pulmonary nodule detection method. The method comprises: (1), CT image pretreatment is carried out, so that the pixel spacing becomes unified and image comparison is unified; (2), a two-dimensional convolution neural network U-net is trained, a pulmonary nodule segmentation image is predicted, and a candidate nodule is recommended based on the pulmonary nodule segmentation image; and (3), a three-dimensional deep residual neural network Resnet3D is trained, a true-false positive probability of the pulmonary nodule is predicted, and a false positive nodule is screened out. According to the detection method disclosed by the invention, the deep learning advantages are utilized fully, so that the pulmonary nodule can be detected in a CT image automatically, efficiently and accurately; and the adaptability to medical big data is high.

A video inspection system and method for facilitating inspection of a rail component while traveling on the railroad track. The system includes a light source that provides illumination to a rail of the railroad track, a triggering device for automatically providing a trigger signal, a camera adapted to provide an image of the illuminated rail component, and a computing device adapted to capture the image provided by the camera based on the trigger signal. A method for inspecting rail components is also provided, the method including the steps of illuminating a rail of the railroad track, automatically providing a trigger signal, providing a camera adapted to provide an image of the rail component, and capturing the image of the rail component that is provided by the camera based on the trigger signal.

A system and method for fraud detection and management are provided. The system includes a first communication device that receives a phone call from a second communication device, wherein a call flow of the phone call comprises one or more distinct phases. The system also includes a fraud detection and management system (FDMS) platform that determines whether the phone call exceeds a predetermined risk threshold at each distinct phase of the call flow.

The present invention relates to a real time night time vehicle detection and identification system comprising of an illuminant object image segmentation device 1, an illuminant object classifying device 2, a vehicle lighting object identification device 3, a vehicle position determining device 4 and a vehicle tracking device 5. Under various circumstances of road lighting during nighttime, the system can efficiently and accurately demarcate and identify the lamps of incoming and preceding vehicles and accurately provides the driver with auxiliary information needed to analyze the traffic conditions in front of the vehicle during the conditions met on the road at that time.

A displacement sensor having a rectangular shaped elastic member. A piezoelectric element is attached to a first main face of the elastic member. The piezoelectric element has a rectangular-shaped piezoelectric sheet and electrodes on both main faces of the piezoelectric sheet. The piezoelectric sheet is made of poly-L-lactic acid and is at least uniaxially-stretched. The piezoelectric element is attached so that the uniaxial-stretching direction of the piezoelectric sheet is 45° relative to a long-side direction of the elastic member. When the elastic member is bent along the long-side direction, the piezoelectric sheet is stretched along the long-side direction, and the piezoelectric element generates voltage of predetermined level.

A method for automatically producing video cartoons with user-provided faces superimposed on cartoon characters is presented. A cartoon template stores the original video cartoon with unanimously-coloured backgrounds and original cartoon characters' faces drawn following the guidelines of the invention. A background detection algorithm is used to detect the unanimously-coloured backgrounds, while a face detection algorithm is used to detect the facial colours, facial expressions, sizes, rotations, positions and clippings of the original cartoon characters' faces. The detected unanimously-coloured backgrounds may be replaced by optional user-provided backgrounds. User-provided faces with the best matching user-specified facial colours and expressions are superimposed over and on top of the corresponding original faces of the cartoon characters at the detected sizes, rotations, positions and clippings to produce the final video cartoon. The cartoon template can be distributed to many different users. Different users can add different user-provided faces, background images and audio to the same cartoon template to produce many different final video cartoons.

The invention relates to the field of video monitoring and particularly relates to a flame detection method based on video analysis. The flame detection method comprises the following steps: extracting a binary image of a foreground according to a background in a scene image and getting a plurality of foreground object images according to the binary image of the foreground; and screening according to the color feature of a red, green and blue (RGB) histogram of the plurality of the foreground object images and the scale invariant feature transformation (SIFT) of the plurality of the foreground object images or the speeded up robust feature (SURF) so as to screen out the foreground object images which are in line with the flame feature. Through the flame detection method based on the video analysis, which is provided by the embodiment of the invention, the flame situation can be known more efficiently and more accurately.

The invention discloses a method for detecting web page horse hanging. The method is applied to detecting whether web pages contain dangerous data or not, and comprises the following steps of: acquiring web page contents; statically analyzing the acquired web page contents so as to detect web page objects which possibly contain the dangerous data; acquiring the object contents of the detected web page objects; and determining whether the object contents contain the dangerous data or not. The invention also discloses a corresponding device for detecting the web page horse hanging. The method and the system for detecting the web page horse hanging can efficiently and accurately detect whether horses are hung on the web pages or not by statically analyzing the web page contents and dynamically explaining the object contents which are executed dubiously.

The present invention is intended to devise a process of measuring β-amyloid in a biological sample such as blood and to apply the process to diagnosis of Alzheimer's disease.It is possible to assay Alzheimer's disease by measuring a total amount of β-amyloid 1-42 and β-amyloid 1-42 fragments each of which retains a C-terminal site of the β-amyloid 1-42 in the biological sample by an immunological assay in which an antibody which recognizes the C-terminal site of β-amyloid 1-42 is used. It is preferable that the immunological assay be a competitive immunological assay.

The invention discloses a method for detecting a designated location point of variation. The method comprises the following steps that a designated location point of variation and a reference sequence containing the variation are determined based on known information of the variation; sequencing data of nucleic acid of a sample to be detected is obtained, wherein the sequencing data comprises multiple read segments; the read segment containing the designated location point in the sequencing data is extracted, and then the designated read segment is obtained; N bps are extended in the directions of both ends centered on the designated location point of the designated read segment, a designated fragment is obtained, 4<=N<=10; the designated fragment is compared with the reference sequence containing the variation, and then a support read segment is obtained, wherein the support read segment is a read segment where the designated fragment matched with the reference sequence containing the variation is located; the amount of the support read segment is counted, and whether or not the variation exists is judged based on the amount of the support read segment. Attention is paid to whether or not sequence characteristics which should occur after the variation exist in the read segment, and based on the judgment, detection of the designated location point of the variation is carried out; the problem of the decline in the quality of the comparison near the variation location point is avoided, and the variation can be detected quickly and accurately.

A defect detecting apparatus captures images of a protein chip formed on each die of a wafer at a plurality of different focal positions, with respect to every division region obtained by dividing each die in plurality; stores inspection target images for every division region at every focal position together with an ID for identifying each division region; creates a model image for every division region at every focal position by calculating an average luminance value of pixels of each inspection target image having the corresponding ID; extracts a difference between the model image and each inspection target image as a difference image; extracts a Blob having an area larger than a preset value from each difference image as a defect; and classifies the kind of the defect based on a feature point of the extracted Blob.

The invention provides a method and a system for detecting abnormal tracks in driver driving tracks. By comparing with historical driving tracks of a driver and driving tracks of other drivers, whether tracks are abnormal driving tracks is detected. The method comprises the steps of acquiring an original data set, performing preliminary processing to data, performing training to generate an abnormal driving track detection model, and detecting the abnormal tracks in the original data of the driver by using the generated abnormal track detection model to obtain the abnormal driving tracks. The detection system comprises an acquisition device and a detection device. As compared with the existing abnormal driver track detection method, the method is applicable to diversified driver track data, the robustness to the number of abnormal tracks is very good, the abnormal driver tracks can be more accurately and efficiently detected, and better aids can be provided for identity authentication in settlement of insurance claim and identification in personalized service.

The invention discloses a background and inter-frame difference algorithm-based traffic flow and motion state detection method, which comprises the following steps of: first performing gray processing on a real-time road surface image shot by a high-definition monitoring camera; then performing background differential operation on a group of gray-processed image comprising a plurality of frames to obtain a differential image, extracting vehicle characteristic areas, and calculating a traffic flow area and the ratio of the traffic flow area to a road surface area to obtain traffic flow information; next selecting differential images processed by the steps (1) and (2) at moments K and K+1, and performing inter-frame differential operation to obtain a traffic flow area variation and the ratio of the traffic flow area variation to the road surface area to further obtain traffic flow motion state information; and finally comprehensively analyzing the traffic flow information and the motion state information, encoding different road traffic flow conditions, and predicting a road traffic flow condition. Data support can be provided for intelligent traffic lights, a foundation can be laid for the development of intelligent traffic, and the method has broad application prospect.

The invention discloses a method for performing genome amplification by using embryo blastula-stage cells, performing chromosome detection on the preimplantational embryo by combining a high-flux sequencing technique and screening out the chromosome normal embryo. The method can comprehensively and completely analyze the genetic variation information of the embryo genome, thereby instructing the preimplantational embryo selection, reducing the hereditary diseases and enhancing the success rate of test tube babies. The method comprises the following steps: blastula-stage trophocyte separation; genome amplification; DNA (deoxyribonucleic acid) segmentation; and Proton library establishment, mounting sequencing and sequencing data analysis. By using the blastula-stage embryo to perform trophocyte separation detection, the method avoids the injuries of cleavage-stage cell separation to the embryo, obtains higher cell quantity than the cleavage stage, and enhances the success rate and amplification effect of genome amplification. After the blastula-stage embryos are subjected to the natural elimination process, the high-quality blastula-stage embryo is selected for detection, thereby saving the cost.

A system and method for efficiently and accurately detecting errors in a network is provided. Devices coupled to a network are configured to determine their local network status and transmit their local network status to a centralized location. The centralized location is configured to receive the local network status from each device and provide network status for the entire network. The centralized location is also configured to analyze the collective network status information and report any network errors. The centralized location is further configured to present the information in a format that allows a user of the network to quickly determine the configuration and status of the network (such as a computer generated graphical representation of the network).

A biochip for detecting the pathogenic funguses features that the DNA probes respectively for detecting one of 20 pathogenic funguses including Candida albicans, Aspergillus fumigatus, etc are immobilized on the glass plate, silicon chip, or high-molecular material.

There is provided a pattern inspection apparatus that is capable of detecting a defect accurately and efficiently to inspect a pattern of a semiconductor device. The pattern inspection apparatus includes: a contour extraction means for extracting contour data of a pattern from a captured image of the semiconductor device; a non-linear part extraction means for extracting a non-linear part from the contour data; an angular part extraction means for extracting an angular part of a pattern from design data of the semiconductor device; and a defect detection section that compares a position of the non-linear part extracted by the non-linear part extraction section with a position of the angular part extracted by the angular part extraction section so as to detect a position of a defective part of a pattern.

Light from a light source device is polarized through a polarizer and is caused to impinge obliquely onto an object to be inspected. The resulting scattered light is received by a CCD imaging device having an element for separating scattered polarized light disposed in a dark field. Component light intensities are worked out for an obtained P-polarized component image and an obtained S-polarized component image and a polarization direction is determined as a ratio of them. The component light intensities and the polarization directions are determined from images obtained by imaging of the light scattering entities in a state where static stress is not applied to the object to the inspected and in a state where static load is applied thereto so as to generate tensional stress on the side irradiated by light. The component light intensities and the polarization directions are compared with predetermined threshold values.

The invention discloses a high-throughput detection method for grain form parameters. The high-throughput detection method includes the steps: setting up a grain form detection platform; calibrating a detection environment; acquiring and segmenting an image; extracting a contour and acquiring characteristic parameters; sieving adhered grains and impurities by the aid of a grain form parameter preset threshold; analyzing an image in the contour and realizing high-throughput color and texture characteristic detection; outputting grain form parameter detection results. The detection results are accurate and reliable, and compared with a manual measuring method, the grain form parameter detection method greatly improves detection efficiency and stability. The detection requirement for the grain form parameters in gene research can be met, the method is simple to implement, detection is efficient and accurate, and application requirements are met.

The invention provides a detection system and method for product surface defects. The system comprises a grasping mechanical hand, a detection mechanical hand, a control module and a defect detectiondevice, wherein the defect detection device comprises a first camera and a second camera; the first camera is used for shooting an image of the upper surface of a detected product; the second camera is used for shooting the image of the lower surface of the detected product; the first camera sends the collected image of the upper surface of the product to the control module; the control module detects whether the upper surface of the product has a defect or not according to the image of the upper surface of the product, and outputs a first detection result; the second camera sends the collected image of the lower surface of the product to the control module; and the control module detects whether the lower surface of the product has a defect or not according to the image of the lower surface of the product, and outputs a second detection result. Therefore, the automatic detection on the product surface defects can be realized.

The invention discloses an ultrasonic rail flaw-detection trolley system based on deep learning and a control method thereof. The ultrasonic rail flaw-detection trolley system comprises a flaw-detection system, a trolley controller and a first industrial personal computer, wherein the flaw-detection system comprises a second industrial personal computer, a spraying mechanism, a coating mechanism and an ultrasonic detection wheel connected with the second industrial personal computer; the ultrasonic detection wheel comprises a rubber wheel filled with a coupling agent and an ultrasonic transducer arranged on the rubber wheel; the trolley controller is used for controlling the trolley to move along the rail or brake the trolley and is respectively connected with the first industrial personal computer, the second industrial personal computer, the spraying mechanism and the coating mechanism. A modularized design method is applied to the trolley structure, so that the quick detaching of important parts of the trolley can be realized. The invention also applies the deep learning to the rail flaw-detection application, so that the real-time, efficient and high-accuracy detection for the flaw-detection data can be realized.

The present invention discloses a method for automatically recognizing white blood cells in leucorrhea based on the convolution neural network, and belongs to the field of automatically recognizing medical microscopic images by using a machine vision scheme. The method comprises: firstly artificially identifying a plurality of white blood cell images and non-white blood cell images; establishing a 9-layer neural network, and training the 9-layer neural network according to the artificially identified images; and carrying out corresponding adjustment in real time on training parameters and a learning rate of the neural network during the training process; and after completion of training the neural network, carrying out detection on to-be-detected target images. According to the method disclosed by the present invention, effects of reducing labor intensity of image reading by doctors and improving the diagnostic accuracy are achieved.

A sorting apparatus 10 which uses a sorting / conveying line 11 with branch destination, including a branch line 22 for sorting the commodities more than a quantity necessary for the branch destination and a surplus article returning line 23 for returning surplus commodities after removing a required quantity to the sorting / conveying line 11.

Identifying characteristics integrated in a totem and detected by a capacitive display mat are applied to perform functions in response to end user inputs. Identifying characteristics include capacitive feet integrated in a totem bottom surface with identifying features including varying capacitive values for the feet, varying pad sizes and shapes, varying pad patterns and varying capacitive oscillation frequencies. In one embodiment, the surface area of the totem in contact with the capacitive mat varies depending upon the pressure placed on the totem to alter the shape of a compressible material disposed at the bottom surface.

The invention discloses a Java Web framework vulnerability attack general detection and positioning method and system, and the system comprises: a context information collection module which is used for deploying a probe code of an RASP at each key level of an application program, and obtaining key context data such as parameter information of a related sensitive function; a vulnerability judgment module which is used for constructing an attack detection rule matching algorithm according to the established vulnerability utilization rule set, namely the representation characteristics of different vulnerability attack types in each level of the Web program; and a vulnerability alarm module which is used for receiving a judgment result of the vulnerability judgment module and outputting feedback information which comprises a judgment result of vulnerability attacks suffered by the program. The collected context information application rules in the program can be directly matched, meanwhile, the detection rule set can be expanded along with expansion of vulnerability attack types analyzed in the early stage, the operation state in the program can be monitored in real time, unknown vulnerability defense is achieved, and effectiveness and practicability are achieved.

Low complexity edge detection and DCT type selection method to improve the visual quality of H.264 / AVC encoded video sequence is described. Encoding-generated information is reused to detect an edge macroblock. Variance and Mean Absolute Difference (MAD) of one macroblock shows a certain relationship that is able to be used to differentiate the edge macroblock and the non-edge macroblock. Also, the variance difference of neighbor macroblocks provides a hint for edge existence. Then, a block-based edge detection method uses this information. To determine the DCT type for each block, the detected edges are differentiated as visual obvious edge, texture-like edge, soft edge and strong edge. 8×8 DCT is used for texture-like edges and the 4×4 DCT is used for all the other edges. The result is an efficient and accurate edge detection and transform selection method.

The invention provides a method and equipment for detecting fingers in images. The method includes: extracting a candidate finger-palm center axis traversing candidate fingers and a candidate palm from a binary image generated from an image; generating shape features of the candidate fingers and the candidate palm corresponding to the candidate finger-palm center axis; and judging whether the candidate finger-palm center axis stems from a human hand or not according to shape features of fingers and a palm of a human. By the method and the equipment for detecting the fingers in the images, the features with notable physical meanings are extracted from the images, and the fingers are detected by the aid of the extracted features, so that the fingers are detected accurately and efficiently.

In a playback apparatus, when a transmission finished notification for kth decompressed digital data is received via a communication portion, a control portion detects a silence start position in the kth decompressed digital data that is stored in the first buffer, transfers a section of the kth decompressed digital data that is before the silence start position to the second buffer, temporarily stores (k+1)th decompressed digital data that is received via the communication portion in the first buffer following the transfer of the section of the kth decompressed digital data, detects a silence end position in the (k+1)th decompressed digital data that is stored in the first buffer, and transfers a section of the (k+1)th decompressed digital data that is after the silence end position to the second buffer.