33results about How to "Classification is flexible" patented technology

A 3 fiber line switched ring (3FLSR) provides protection for (optical) transmission networking wherein N nodes are connected via 3 lines (or optical fibers) in a ring topology. Two of the three fibers form a main transmission line and may transmit in one direction in the ring whereas the third might transmit in the opposite direction. This architecture of the 3FLSR provides a dual protection scheme, including a bi-directional line switching coupled with a unidirectional line switching. Traffic is categorized based on the level of protection available for the particular traffic type. The first two transmitting fibers form a bi-directional ring, carrying primary and secondary traffic which can survive 2 and 1 failures on the ring respectively. The third fiber may form a unidirectional ring, carrying additional traffic that can be pre-empted in case of multiple failures in the bi-directional ring. The 3FLSR enables reconfiguration of existing 2 / 4 fiber rings and conforms to applicable standards.

The invention discloses a data packet classification method, which includes CPU preprocessing rule set and GPU execution data packet classification. run. Preprocessing operations are done on the CPU. The packet classification operation is the core part of the algorithm, which classifies the input data packets and outputs the rule number with the highest priority matched by the data packets. Packet classification operations are done on the GPU. The invention makes full use of the hardware characteristics and advantages of the GPU, and has greater performance advantages than conventional data packet classification algorithms.

A method for shape classification of an object is provided. Shape categories are provided which specify a plane and points therein relative to the object, and also specify at least one limit coordinate for each such point, the limit coordinate defining a boundary in a direction normal to the plane for the shape of the object considered in order for the object to be classified into a respective shape category. The shape categories can be provided by a user, making the method very flexible. The shape categories can in particular be derived from a set of samples of objects representing a shape category to be defined. For classification, the position of a surface of the object is measured at each of the points defined in the shape category, and the result is compared with the corresponding limit coordinate.

The invention discloses a vehicle-mounted ad-hoc network intersection prediction routing method based on a CP (Counter Propagation) neural network, and mainly aims to solve the problems of frequent link fracture among nodes in a data packet forwarding section, large data packet transmission delay and low transmission success rate due to high-speed moving of vehicles, frequent change of network topology and a plurality of vehicle distribution densities in a traffic environment. According to the implementation scheme, a next section priority at which data packets are forwarded on an intersection is classified by use of the CP neural network; when a node arrives at the intersection, an average link communication probability of adjacent sections, average node densities of the adjacent sections and a distance ratio of a distance from a next intersection to a destination to a distance from the current intersection to the destination are taken as inputs of the CP neural network in one Hello message period; the priorities of adjacent data packet forwarding sections are taken as outputs; and a section with a highest priority is selected to be an optimal data packet forwarding section. Through adoption of the vehicle-mounted ad-hoc network intersection prediction routing method, the data packet forwarding success rate is increased, and the transmission delay of the data packets is shortened. The method can be applied to a vehicle-mounted ad-hoc network.

The invention discloses a method and device for classifying pictures. Wherein, the method includes: determining the classification basis of the target picture; extracting the picture features of the target picture from the target picture according to the determined classification basis, wherein the picture features include classification features of at least two dimensions belonging to the determined classification basis; The extracted image features classify the target image. The invention solves the technical problem that the classification results obtained by classifying the target pictures have large errors because the related technology adopts a fixed classification basis and a single classification feature is used to classify the target pictures during specific classification.

A method for shape classification of an object is provided. Shape categories are provided which specify a plane and points therein relative to the object, and also specify at least one limit coordinate for each such point, the limit coordinate defining a boundary in a direction normal to the plane for the shape of the object considered in order for the object to be classified into a respective shape category. The shape categories can be provided by a user, making the method very flexible. The shape categories can in particular be derived from a set of samples of objects representing a shape category to be defined. For classification, the position of a surface of the object is measured at each of the points defined in the shape category, and the result is compared with the corresponding limit coordinate.