44results about How to "Good discernment" patented technology

A scintillator array has a compact, two-dimensional arrangement of scintillators partitioned by light reflecting elements and light transmitting elements interposed between the scintillators. That is, the scintillator array has a lattice frame in the form of a crosswise combination of plate-like optical elements such as light reflecting elements and light transmitting elements. The lattice frame defines a plurality of cubicles. A transparent adhesive is poured into a rectangular vessel for receiving the lattice frame, the lattice frame is placed in the rectangular vessel, and then the scintillators are placed in the rectangular vessel. A cured resin object integrating the optical adhesive having cured, the lattice frame and the scintillators is removed from the vessel, and is contoured to make the scintillator array. In this way, the optical elements are arranged in the scintillator array without cutting them with a dicing saw or wire saw.

The invention relates to a shielded workpiece identifying method and device based on binary system feature matching. The identifying problem of a shielded workpiece is solved, and meanwhile the computation complexity can be lowered, the storage space can be reduced, and the matching precision can be improved. In addition, good robustness is achieved on ambient light, view angle changing and partial shielding, the shielded workpiece can be identified in different interference environments, and a good identifying result is obtained. Thus, the shielded workpiece identifying method and device can be widely applied to the field of image identifying.

The invention relates to a network live video feature extraction method in a complex scene based on joint attention ResNeSt. Firstly, key frame extraction is performed on a network live video to obtain key frame data of the video; in order to utilize multi-scale features of video frames, a parallel path is designed according to a multi-scale structure of a feature pyramid network. The parallel path is constructed from bottom to top, information exchange is carried out between the parallel path and an original main path by utilizing transverse connection and oblique connection, and the transverse connection and the oblique connection are convolution operations. Considering that the picture representation form of network live broadcast is mostly a human main body, and a large amount of redundant information is mingled, the space-channel joint attention is introduced, and the picture main body characteristics are conveniently focused. And finally, a ResNeSt feature extraction module is constructed by combining the parallel feature pyramid fused with the joint attention with a convolution layer and a pooling layer, and feature extraction of the network live video in a complex scene is realized through superposition of multiple layers of modules.

The invention discloses a three-dimensional ear recognition method based on block statistic features and dictionary learning sparse representation classification. A three-dimensional ear sampling image is equally divided into a plurality of sub regions; as for each sub region, an ear surface type is firstly calculated, a histogram is then used for counting the ear surface type in each sub region, surface type histograms of all sub regions are finally spliced to serve as a feature description operator for the ear depth image, a dictionary learning sparse representation framework is used for classification, and thus the recognition efficiency and the accuracy are improved. The method of the invention can be used in situations with strict requirements on identity recognition, and can effectively solve the alignment deviation problem among multiple three-dimensional ear sampling.

The invention discloses an effective training method based on PCB noise annotation data, and the method comprises the following steps: S1, collecting images, and building a PCB image data set containing noise annotation and a PCB image data set after data cleaning; S2, respectively extracting image features of each image in the two data sets through an integrated Effective net model; S3, performing dimension reduction on the extracted image features through a deep convolutional network; S4, training a classifier model by adopting a cosine loss function; and S5, alternately training the two data sets by adopting transfer learning, and carrying out fine adjustment on a relatively high-resolution image in the PCB image data set containing the noise annotations to finally realize data training. The invention is mainly used for solving the problem of realizing image retrieval when the training set contains a large number of noise annotations. Efficient speed and accuracy are realized in theretrieval process.

The invention provides a male chronic prostatitis/chronic pelvic pain syndrome pain severity prediction model and establishment thereof, and relates to the field of chronic prostatitis/chronic pelvicpain syndrome pain severity prediction. The pain severity prediction model is established by adopting an advantage ratio OR of variables such as the age of a patient, the level of lecithin bodies in prostatic fluid, urine suppression, anxiety or irritability, contraception, smoking and the like, establishing a column diagram as a model to predict the pain severity of the CP/CPPS patient on the basis of the confidence intervals of 2.5% and 97.5% and the P-value. Establishment of the model mainly comprises the steps of material selection, grouping, variable screening, variable analysis and the like. According to the method, the defects in the prior art are overcome, the pain level of prostatitis can be accurately predicted by establishing the model, and meanwhile, the clinical decision-making efficiency is improved.

A method for performing cellular classification includes extracting a plurality of local feature descriptors (220) from a set of input images (210) and applying a coding process to convert each of the plurality of local feature descriptors into a multi-dimensional code (225). A feature pooling operation (230) is applied on each of the plurality of local feature descriptors to yield a plurality of image representations and each image representation is classified as one of a plurality of cell types (240).

The invention discloses a deep fusion model applied to mammary gland X-ray image anomaly recognition and a classification method of the deep fusion model, and relates to the technical field of image processing methods.The method comprises the steps that a mammary gland X-ray image is preprocessed to remove noise and improve image quality; collecting regions of interest (ROI) of abnormal tissue image categories (i.e. Benign and malignant), and extracting the ROI from random positions of normal tissue image categories; randomly extracting a smaller ROI sub-block image from the ROI image; constructing a depth fusion model, and training the depth fusion model by adopting the ROI sub-block images; finishing classification of each ROI image by using majority voting; according to the method, thedepth information of all the five pre-trained VGG16 blocks is fused, so that the information between different channels of each module is highly correlated, a 1 * 1 convolution layer integrates cross-channel information and further realizes dimensionality reduction, and model parameters can be effectively reduced, so that the performance of the model is improved.

The invention provides a voice signal feature learning method based on a first derivative of Mel-spectrogram (DMS). The method comprises inputting a disease voice sample and a healthy voice sample based on data driving; splitting all samples into frames to extract the DMS versus time, determining the training sets and the test sets of the disease sample and healthy sample by a cross-validation method; training dictionaries for healthy voice and pathological voice separately by using a clustering algorithm, subjecting the DMS of each sample in the two training sets and the two test sets to linear coding and to pooling by using a minimum pooling method to obtain the final features. The supervised method makes full use of tag information, and the learned features have better discriminating power.

The invention provides a high-precision crop pest and disease damage identification method, which comprises the following steps of: S1, enabling a user to input a crop leaf with any size, and scalingthe crop leaf to a uniform size; S2, converting the picture obtained in the step S1 from an RGB channel to a YCrCb color space; S3, merging the YCrCb color space three-channel picture obtained in thestep S2 into an original RGB space to form six-channel input, and then sending the six-channel input into a network through corresponding normalization processing; S4, sending the data obtained in thestep S3 into a network structure proposed by the design, and training to obtain a prediction classification category and a saliency map. The invention belongs to the field of computer vision application, considers the workload and specialty of crop disease and insect pest identification, uses a deep learning technology to replace traditional manual work to greatly reduce the cost, has the advantages of high precision, high speed and the like, and can deploy a model at mobile terminals such as a mobile phone, a tablet personal computer and the like offline to facilitate the use of a user.

The invention belongs to the technical field of communication, and particularly relates to a device and method for reflecting a prompt event through flashing of a prompt lamp and a mobile terminal. The method for reflecting the prompt event through flashing of the prompt lamp comprises the following steps: when the prompt event is obtained, judging type information of the prompt event, according to the type information of the prompt event, obtaining a control instruction for flashing of a prompt lamp, and issuing a control instruction to the prompt lamp so that some important notifications inthe mobile phone can be mastered through the notifications of the indicating lamp, when the mobile phone has an information notification prompt, the prompt lamp carries out flashing prompt in different flashing frequencies, brightness, colors and the like according to different settings of the notification, so that the notification content of the mobile phone can be identified without lightening ascreen, an excellent identification effect is achieved, and the experience feeling of a user is enhanced.

The invention provides memory module automatic testing equipment. The equipment comprises a rack platform and at least one group of testing mechanisms arranged on the rack platform, wherein each testing mechanism comprises a manipulator, an X-direction moving carrying station, a Y-direction moving carrying station, an X-direction material box module, a Y-direction material box module, a labeling machine, a code scanning gun, an ion fan and a memory testing instrument; the X-direction material box module corresponds to the underside of the X-direction moving carrying station, and comprises a first X-direction material box group, a second X-direction material box group and a third X-direction material box group, wherein the third X-direction material box group and the Y-direction material box module are linearly and correspondingly distributed under the Y-direction moving carrying station; the Y-direction material box module comprises a first Y-direction material box group, a second Y-direction material box group and a third Y-direction material box group, and the manipulator, the labeling machine, the code scanning gun, the ion fan and the memory testing instrument are arranged on the right-angle inner side of the X-direction material box module and the Y-direction material box module. The equipment can greatly improve the automation degree of the memory module, thus greatly improving the testing efficiency.

The invention relates to an extraction method and application of salt-tolerant yeast, the part of 5-10cm below unstirred soy sauce mash in a fermentation tank is used as a sample, a culture medium A and a culture medium B are prepared, the culture medium A comprises 1-10g of glucose, 0.1-2g of yeast powder, 0.1-1g of polypeptone, 0.1-2g of monopotassium phosphate, 1-8g of agar powder and 100g of water, the culture medium B is a mixture of raw soy sauce, salt and water, and the mixture is mixed with the raw soy sauce, the salt and the water. Uniformly mixing A and B, adding a sample to prepare a suspension, and performing plate separation and purification to obtain the salt-tolerant yeast; according to the method, the saccharomycetes are extracted, and after the saccharomycetes are subjected to enlarged culture, the generated alcohol is good in aroma, accompanied with special fragrant and sweet taste, and has discriminative taste; the salt-tolerant yeast is subjected to enlarged culture and added into soy sauce mash, and participates in the fermentation process as dominant bacteria, so that the beany flavor of the soy sauce can be effectively reduced, the special fragrance of the salt-tolerant yeast can enrich the flavor of the soy sauce, and the soy sauce has better discrimination than other soy sauce.

The invention discloses a lightweight iris recognition system for a mobile terminal. The system comprises a preprocessing module for performing smoothing processing, lens detection and white hole filling on an iris image captured by an infrared camera; an iris positioning module used for carrying out coarse positioning and accurate positioning on the iris image, obtaining a circle center and a radius of a pupil through coarse positioning, obtaining inner and outer boundary points of an iris through accurate positioning, realizing iris region normalization through fitting a circle by the boundary points, and recording the position of noise by adopting a mask; a coding module used for coding the iris image, carrying out feature extraction on the normalized iris area by adopting a Gabor filter to obtain a feature map, and selecting a point at a specified position in the feature map for coding; a registration and recognition module used for registering the iris image, binding iris codes, masks and registrant information and storing the iris codes, the masks and the registrant information in a database, and used for recognizing the iris image. The system solves the problems of low irisrecognition precision and large resource consumption of the mobile terminal.