284results about How to "Reduce fit" patented technology

This invention relates to polyolefin compositions. In particular, the invention pertains to elastic polymer compositions that can be more easily processed on cast film lines, extrusion lamination or coating lines. The compositions of the present invention preferably comprise an elastomeric polyolefin resin and a high pressure low density type resin.

This invention relates to polyolefin compositions. In particular, the invention pertains to elastic polymer compositions that can be more easily processed on cast film lines, extrusion lamination or coating lines. The compositions of the present invention preferably comprise an elastomeric polyolefin resin and a high pressure low density type resin.

The invention provides a peelable glue coating composition and a peelable glue coating as well as a preparation method and an application of the peelable glue coating. The peelable glue coating composition comprises the following components in parts by weight: 80-120 parts of an acrylate type oligomer, 8-20 parts of a dilutive monomer, 0.1-1.5 parts of a peeling force modifying agent, 2-12 parts of a triggering agent and 30-50 parts of an organic solvent. The peelable glue coating composition is relatively high in transparency and relatively high in light transmittance and relatively low in haze after film formation. Meanwhile, a coating formed by curing the peelable glue coating composition is relatively high in high temperature resistance and relatively high in peelability. By the high temperature resistance, the coating is good in adhesion property on the surface of a functional hardened layer, and the problems of air bubbles and warping can be effectively improved. By the peelability, the problem of glue layer residues caused during peeling of the coating can be reduced.

The invention provides a landslide displacement prediction method based on wavelet transform-rough set-support vector regression (WT-RS-SVR) combination. By means of the method, according to the characteristics of the influence factors of landslide displacement, the complex displacement process and landslide displacement monitor data measured in real time, accumulative displacement of a typical monitor point is decomposed into trend-term displacement and periodic-term displacement through WT, and a trend-term displacement prediction function is obtained through curve fitting; screening is conducted on the influence factors of the landslide displacement through an RS algorithm, and selected factor sets are used as input factor sets of an SVR machine, accordingly a landslide displacement optimization prediction model based on WT-RS-SVR combination is established, and the precision of a prediction result is analyzed and evaluated. The prediction result of the landslide displacement prediction method can well embody the development and change tendency of the landslide displacement. The landslide displacement prediction method has high prediction capacity, and is accurate, effective and practical.

The invention relates to a medical image sample screening method and device, computer equipment and a storage medium, for carrying out intelligent screening on unlabeled medical image samples. The medical image sample screening method includes the steps: carrying out model training on a labeled sample set by utilizing a Mask-RCNN model so as to obtain a focus target detection depth model; predicting the unlabeled medical image sample set according to the focus target detection depth model to obtain a prediction result of each medical image sample and judge a labeling value; and selecting a medical image sample with high annotation value to perform annotation confirmation, performing iterative updating on the focus target detection depth model, and ending the iterative updating until the performance of the focus target detection depth model cannot continue to annotate a new sample. According to the medical image sample screening method, under the condition of limited computing resourcesor annotation cost, the high-value small data set is actively mined and extracted, and efficient diagnosis and decision making are achieved by simulating a medical expert intelligent learning mode, and the intelligent degree is high, and the processing speed is high, and the problem that the annotation efficiency is low is effectively solved.

The invention relates to the technical field of display, and provides a flexible display panel and a preparation method thereof, a display device and a display module. The flexible display panel comprises a plane area, a first bending area, a second bending area and a corner bending area, the first bending area is connected to the first edge of the plane area in the first direction, the second bending area is connected to the second edge of the plane area in the second direction, and a set angle is formed between the second direction and the first direction; a corner bending area is connectedto the corner of the plane area; a corner bending area is arranged on the base and connected between the first bending area and the second bending area, the corner bending area is provided with a plurality of first hole sets, after the corner bending area is bent, the width of the first hole set located in the middle of the corner bending area is increased, and the width of the first hole set located in the edge area of the corner bending area is decreased. Tensile strain and compressive strain are absorbed through the first hole set, and attaching cracks and edge area wrinkles of the flexibledisplay panel are reduced. No bubbles can be generated.

This invention discloses a high penetration light-increasing optical component including forming a LC film of the cholesterin phase and an 1/4 wavelength delay film on a matrix orderly and further including forming a polarized film on the delay film, in which, the first two films are formed by coating and the polarized film is formed by coating or adhering a pre-formed polarized film on the delay film. This invention with the polarized film combines the high penetration light increasing optical component with the backlight module of the LED with the polarized film as the brightness increased polarized light source of a LED.

The invention provides a construction method for a network connection adaptive deep convolutional model. The method comprises the steps that first, weight vectors in a convolutional neural network areorthogonalized; second, connection between layers in the convolutional neural network is deleted according to a norm; and third, an activation function is constructed. According to the method, data relevancy is lowered through weight vector orthogonalization; the connection between layers in the convolutional neural network is deleted based on the p norm, the problem of overfitting in the construction process of the deep convolutional model is effectively lowered, and network connection adaptation is realized; and the activation function is constructed to generalize data information as much as possible on the premise of not changing a parameter scale, and therefore the accuracy of the deep convolutional model is improved.

The invention provides a 3D lung nodule generation method, a device and an electronic device, which relate to the technical field of image recognition. The method comprises the following steps: obtaining target lung nodule image data; inputting the target lung nodule image data into the 3D lung nodule generation model, and outputting the 3D lung nodule corresponding to the target lung nodule imagedata; the 3D lung nodule generation model being trained by depth convolution based on Wasserstein distance to generate the real nodule data by DCGAN; the real nodule data being obtained by data enhancement processing of lung nodule image data to be trained; a depth convolution based on Wasserstein distance being used to generate a countermeasure network DCGAN to train the real nodule data, wherein, the real nodule data is obtained by data enhancement processing of the lung nodule image data to be trained, which increases the amount of data of model training, improves the model accuracy and reduces the model over-fitting, so that the model has high accuracy in the real application.

The invention discloses an integrated pouring device for installation of a rail column type inspection pit fastening system. The periphery of a stand column reserved steel bar is wrapped by shaping aluminum formworks, wherein the two shaping aluminum formworks parallel to a steel rail above the reserved steel bar are higher than the other two shaping aluminum formworks at the positions perpendicular to the steel rail, and the shaping aluminum formworks are used for fixing a matched pouring funnel. The invention further discloses a laying method of the integrated pouring device. The integrated pouring device replaces traditional woodworking operation wasting time and materials, solves the problems that the formwork supporting effect is slow and parts are prone to contamination in the pouring process, and is short in construction period, low in use cost, convenient to construct, high in efficiency, good in stability and wide in application range.

The invention relates to a general positioning method for half-ship refloating and pier falling. The method consists of a positioning guide device, a stabilizer device, a limit device and a series ofweight, cable and precision control schemes to form a set of more versatile positioning methods for the ship to refloat and piers to fall., t, the method is not only applicable to the refloating and piers falling of half-ship products, but also applicable to the secondary docking and lifting piers of various offshore platforms. The invention effectively solves the problem of how to safely and efficiently complete the floating and lifting and the secondary lifting piers of ships under various working conditions, reduces the deformation of the ship hull and shortens the docking period.

The invention discloses a cloth defect detection method and system based on a deep neural network, and belongs to the technical field of pattern recognition. The method comprises the steps that a defect cloth image training set is used for training a deep neural network model, labels are defect types and real frame position information, the deep neural network model is composed of a backbone network and a detection network, and the backbone network is used for extracting three feature maps with different scales from defect cloth images; the detection network includesthree detection sub-networks and the detection result fusion module, wherein the three detection sub-networks are the same in structure. Each detection sub-network is used for detecting a defect type and prediction frame position information from the feature map, and consists of three dense connecting blocks, and the feature channel connection between the dense blocks is used for enhancing feature transfer, and the detection result fusion module is used for performing non-maximum suppression on the prediction result to obtain a final prediction frame and a defect type, and inputting to-be-detected cloth into the traineddeep neural network model to obtain a detection result, so that the type and the position of the defect in the cloth can be detected more quickly and accurately.

The invention discloses a fabrication method of a vertical-structure blue-light LED chip. The fabrication method comprises the steps of firstly, taking a substrate with a low stress buffer layer as a growth foundation; secondly, sequentially growing other epitaxial layers on a smooth u-GaN surface to obtain a blue-light LED epitaxial wafer; and finally, fabricating a vertical-structure LED from the blue-light LED epitaxial wafer, wherein the step of fabricating the vertical-structure LED mainly comprises the following steps of depositing and forming a reflecting mirror on a surface of the LED epitaxial wafer, fabricating a metal electrode pattern, bonding a surface of the metal electrode pattern on a metal substrate by a high-temperature metal bonding process, stripping the substrate by a laser lift-off technology, and fabricating another metal electrode pattern on the u-GaN surface. The vertical-structure LED fabricated according to the method is relatively high in single-chip power, the quantity of LEDs in series and in parallel is reduced, the user demand can be satisfied by a single chip, meanwhile, a driving circuit design is simplified, the reliability of an LED product is greatly improved, and the service lifetime of the LED product is greatly prolonged.

The invention discloses a small sample classification model construction method based on transfer learning and an application of a small sample classification model constructed by the method in iris image classification. The method comprises the steps: constructing an ICP-VGG model based on VGG16 model transfer learning; configuring an activation function of a full connection layer and a dropout ratio of a Dropout layer in the self-defined network according to the iris image task; finely tuning the network, and setting model training related parameters; acquiring a small sample iris data set,and performing data preprocessing and data enhancement on the data set; training and verifying the model, and outputting a recognition result image. According to the method provided by the invention,the deep learning model can be better applied to the field of small sample irises, the over-fitting is reduced, and the recognition accuracy is improved.