11results about How to "High classification accuracy" patented technology

Methods and apparatus for generating at least one output data set from at least one input data set for use in association with a data mining process are provided. First, data statistics are constructed from the at least one input data set. Then, an output data set is generated from the data statistics. The output data set differs from the input data set but maintains one or more correlations from within the input data set. The correlations may be the inherent correlations between different dimensions of a multidimensional input data set. A significant amount of information from the input data set may be hidden so that the privacy level of the data mining process may be increased.

The invention relates to a text classification method and device. The method comprises the following steps of: obtaining a to-be-classified text, wherein the to-be-classified text comprises a feature vocabulary; obtaining a classification model and feature weight vectors of a plurality of text categories corresponding to the classification model; calculating a voting score of a text category corresponding to the feature vocabulary according to the feature weight vectors of the plurality of text categories so as to obtain a text category with the highest voting score; and determining the text category with the highest voting score as a text category corresponding to the to-be-classified text. By adopting the method to carry out real-time online classification on the texts, the server resource consumption can be effectively relieved.

A method classifies data into multiple classes so that the data in each class have a class-conditional probability distribution. The class-conditional probability distributions of measured data are projected into a likelihood space. The projected class-conditional probability distributions in the likelihood space are then classified according to a discriminant classifier in likelihood space.

The invention discloses an automatic fry grading system which comprises a transparent pipe, a grading tank connected with the transparent pipe and an industrial personal computer connected with the grading tank. The inlet end of the transparent pipe is connected with a hose which is connected with a pond through a water tank with a slot. The outlet end of the transparent pipe is connected with thegrading tank. The outlet end of the grading tank is connected with a plurality of grading pipes. A CCD camera is arranged on the side face of the transparent pipe and connected with the industrial personal computer. The industrial personal computer is connected with a swing arm in the grading tank. The industrial personal computer is connected with valves at the inlet ends of the grading pipes. The automatic fry grading system can rapidly grade fish and is high in grading efficiency, accurate, automatic in grading process control, high in intelligent degree and good in safety.

The invention discloses an auger-delivery walnut feeding grading device and shell-kernel separating system. A walnut auger delivery device, a circular grid grading device and a three-grade shell-kernel separating system are included. An auger delivery mechanism is arranged at the bottom of a V-shaped feeding hopper and penetrates through the V-shaped feeding hopper and the circular grid grading device; the posture of walnuts is adjusted by the auger delivery mechanism when the walnuts are delivered, thereby facilitating accurate grading of the circular grid grading device; after the walnuts are graded, the circular grid grading device also adjusts the walnut posture to prepare for shell breaking; and each grade of shell-kernel separating device of the three-grade shell-kernel separating system correspondingly separates walnut shells and kernels with varying degrees of sizes. According to the auger-delivery walnut feeding grading device and shell-kernel separating system, grading of thewalnuts with different sizes can be realized, the grading structure is simple, a walnut shell breaking device is cooperatively arranged on the shell-kernel separating system; and the three-stage shell-kernel separating devices correspondingly separate the walnut shells and kernels with different sizes, and the separating effect is good.

The invention discloses a tea tender sprout grade identification and classification method based on computer vision. The method comprises the following steps: acquiring an original RGB image of tea on a tea tree; secondly, preprocessing the original RGB images of the tea leaves, and directly identifying and segmenting tender sprouts of the tea leaves on the tea trees by adopting an improved watershed algorithm; then, according to the recognition result of the improved watershed algorithm, carrying out classified picking and grade classified marking on the tender sprouts of the tea according to one bud and one leaf, one bud and two leaves and one bud and multiple leaves; thirdly, shooting a plurality of three types of sample images of one bud and one leaf, one bud and two leaves and one bud and multiple leaves which are subjected to grade classification marking; and finally, putting the plurality of three types of sample images into a LeNet5 convolutional neural network, and carrying out tea tender leaf grade classification training and testing. The method has the advantage of high tea grade classification accuracy.

A vertical micrometer grader relates to the technical field of material grading. A main body of the vertical micrometer grader comprises a vertical grader, a cyclone separator, a deduster and a draught fan which are sequentially connected in series. The vertical grader is further connected with a charging device, and a star-shaped discharging valve is respectively arranged at the lower ends of the charging device, the vertical grader, the cyclone separator and the deduster. The vertical micrometer grader is reasonable in structural design, can grade spherical particles, sheet-shaped particles and needle-shaped particles, can grade particles with different densities, is high in grading accuracy and grading efficiency, can be connected with various of grinding devices in use to form closed circulating work, and effectively improves work efficiency. In addition, product particle sizes can be adjusted in stepless mode, and variety renovation is very convenient.

The invention provides a tissue culture seedling grading method based on a two-dimensional image and three-dimensional growth information, and belongs to the technical field of medicinal plant cultivation. The method comprises the following steps: performing color space transformation on a plant RGB image acquired in a surrounding manner to obtain an HSV image; fusing a G component in the RGB image and an S component in the HSV image to obtain each plant main body of the tissue culture seedling; shooting a plurality of plant images of each plant in a multi-angle surrounding manner; establishing a deep learning grading model, and performing tissue culture seedling three-dimensional model reconstruction on all the shot RGB images by adopting a motion structure recovery algorithm; and comparing the shape parameters of each plant obtained from the reconstructed three-dimensional model with actual measurement parameters, and grading the tissue culture seedlings according to a grading principle in combination with a grading result of a deep learning grading model. According to the method, the tissue culture seedlings are classified according to the two-dimensional color image and the three-dimensional growth information, so that the classification precision of the tissue culture seedlings is improved, and meanwhile, rapid classification of the tissue culture seedlings is realized.

The invention provides a method and device for structuring and grading a tubular structure blood vessel and electronic equipment, and the method for structuring and grading the tubular structure blood vessel comprises the steps: carrying out the distance transformation of a tubular structure blood vessel image, and obtaining a distance transformation result of each pixel point in the tubular structure blood vessel image; carrying out region growing based on the distance transformation result to obtain a mediastine image, and carrying out subtraction operation on the tubular structure blood vessel image and the mediastine image to obtain an intrapulmonary blood vessel image; extracting a center line of the intrapulmonary blood vessel image; performing line segment extraction based on the center line and a hierarchical line segment hierarchical relationship to obtain a center line structure image; and mapping the midline structure image to the tubular structure blood vessel image to obtain the structured hierarchical tubular structure blood vessel. The grading precision of the tubular structure blood vessel can be improved.

The invention relates to a device used for separating fine powder from powder through air, in particular to a dry type air classifier suitable for classifying non-sticky materials. The dry type air classifier comprises a main machine barrel body. A feed pipe is arranged at the lower part of the main machine barrel body, and a fan opening is formed in one side of the main machine barrel body. The dry type air classifier is characterized in that a skirt type elutriation ring located on the same horizontal line as the fan opening is arranged in the main machine barrel body; a classifying cone is connected to the lower part of the skirt type elutriation ring; a classifying rotor connected with the main machine barrel body is arranged at the upper part of the skirt type elutriation ring; and a fine powder discharging bend installed on the main machine barrel body is arranged on one side of the classifying rotor. The dry type air classifier has the advantages that the classifying range is wide, the classifying accuracy is high, the classifying efficiency is high, the structure is simple and the occupied area is small, is suitable for classifying the non-sticky materials and can carry out dry type classification on various kinds of organic matter and inorganic matter.