167results about How to "Fast retrieval" patented technology

The invention discloses a fast multi-label picture retrieval system and a realization method. The method comprises the following steps: deploying an RPN (Region Proposal Network) for extracting region proposals in a convolutional neural network, extracting region proposal information of pictures, and performing ROI pooling calculation on the region proposal information; after pooling, building a multi-label classification loss function through a fully connected layer according to multi-label information to train the convolutional neural network, and building a weighted three-dimensional loss function to train the convolutional neural network; extracting the hash code of each picture from a picture candidate set through the convolutional neural network after multi-task learning, saving the hash codes to a database, and comparing the hash codes with the hash codes in the database, thus completing picture retrieval. The whole network is trained through multi-task learning of classification and hashing, and therefore, the accuracy of retrieval is ensured. Moreover, the similarity is measured using Hamming distance in the process of retrieval, and the efficiency of retrieval is improved greatly.

The invention relates to a retrieval method for similar pictures, which comprises the following steps that: a receiving module is used for receiving source pictures designated by customers; a setting module is used for setting the retrieval similarity of the similar pictures, wherein the similarity comprises rough retrieval similarity; and a rough retrieval module is used for extracting rough characteristics from the source pictures and the prior pictures, compares the rough characteristics of the source pictures and the prior pictures according to the rough retrieval similarity and searches rough similar pictures. The invention provides a retrieval system for the similar pictures. The retrieval method and the retrieval system can quickly and accurately search the similar pictures.

The invention relates to a knowledge acquiring system field, in particular to a knowledge intelligent acquiring system of vegetable supply chains, aiming at designing a set of knowledge intelligent acquiring system of vegetable supply chains, realizing the effect knowledge spreading of the vegetable supply chains by means of the computer media, and particularly realizing the knowledge transfer in the field of the vegetable supply chains at home and aboard. The system takes the ontology model of the vegetable supply chains as a support and provides three retrieval modes, i.e. keyword retrieval, semantic extension retrieval and semantic retrieval based on the ontology model as a tool for acquiring the knowledge in the professional field. The system has high query efficiency and facilitates the participants of the vegetable supply chains and scientific researchers to acquire the professional knowledge.

The invention discloses a sole trace pattern image retrieval method based on a multi-feedback mechanism, comprising the following steps: step 1, sorting and outputting sole trace pattern images pre-stored in a sample image library according to the degree of similarity between the spectrum characteristics of binary charts of sole trace pattern images acquired on site and the features of the images in the sample image library; step 2, selecting at least two of the output sole trace pattern images pre-stored in the sample image library according to a predetermined standard, averaging the spectrum characteristics of the selected sole trace pattern images according to the spectrum characteristics of the selected sole trace pattern images, and working out the average spectrum characteristic; step 3, calculating the score of similarity between the average spectrum characteristic and the spectrum characteristic of the binary chart of each sole trace pattern image pre-stored in the sample image library; and step 4, outputting the sole trace pattern images pre-stored in the sample image library according to the score of similarity from high to low.

The invention discloses an intelligent skynet rapid video retrieval system which comprises a data collecting and processing module and an intelligence information module. The data collecting and processing module comprises an intelligent DVR and a video abstraction system, wherein the intelligent DVR is connected in a video in real time, intelligent analysis and intelligent detection are carried out on the video by the intelligent DVR, meanwhile intelligent alarm records are generated, the video abstraction system obtains video files, and abstraction compression processing is carried out on the video files by the video abstraction system to form abstraction files. The intelligence information module comprises an intelligence data base, an intelligence information system and a platform client side, wherein the intelligence data base is used for storing the abstraction files, alarm information, original videos and relevant description information which are uploaded from a data collecting and processing layer, the intelligence information system manages all intelligence information in a unified and concentrated mode and provides various inquiry modes for the platform client side, and the platform client side retrieves the uploaded abstraction files and the uploaded alarm records. The intelligent skynet rapid video retrieval system directly and rapidly conducts retrieval according to time, places, channels or label information, repeated video processing is not required, and the intelligent skynet rapid video retrieval system is rapid and effective.

The invention discloses a method for rapidly screening positions from a position database. The method relates to the position database, a position index generator, a position search server and a Web client. The position index generator is interconnected with the position database by an open database. The position index generator generates an index file according to the position data in the position database in an index structure based on a suffix array and distributes the generated index file to the position search server. The position search server is connected with the Web client through a socket. The Web client sends an inquiry request to the position search server through the socket. The position search server searches the index file and returns search results to the Web client for display through the socket. According to the method, all positions in the position database can be searched rapidly and effectively, and the search speed is greatly improved at the premise of guaranteeing the search accuracy.

The present invention discloses a traffic video monitoring system based on standard labeling and quick search, comprising a traffic monitoring video capturing module, a video image content analyzing module, a traffic video key frame obtaining module, a monitoring expansion data encoding module and a data stream searching module. Based on an SVAC standard, the grammar and semantics of the monitor messages can be appropriately expanded to satisfy the requirements of quick search of monitoring videos. The system provides a unified standard interface to realize labeling and quick search of compressed monitoring video data, and the monitor messages can be appropriately stretched and clipped according to business requirement to meet demand of functions such as data transmission. Meanwhile, an encoding scheme targeting traffic objects and events is worked out and the SVAC standard is used to realize a highly-accurate, simple, effective, unified and fast labeling and quick search method of traffic objects and events.

The invention discloses an image and video retrieval method, which comprises the following steps: decoding images of an image source; cutting, converting and zooming a target image; treating a gray-scale map; counting a gray level histogram of the gray-scale map; converting the gray-scale map to a black and white picture; dividing the black and white picture; calculating a sum; storing the sum into a database to form a query range; inquiring, comparing and calculating errors; indexing fields; and inquiring and comparing to obtain a matching result. Through feature induction and accurate calculation for images and video frame images, the method has quick retrieval speed for the images and the video and high retrieval accuracy; and when the image quality, definition, color tones, watermarks and captions are changed, the retrieval accuracy and speed are not influenced, so the retrieval process is quick, the method is simple and the effects are obvious.

The invention discloses a method for obtaining personnel information of buyers. The method comprises the following steps that comment information of the buyer on a commodity in the historical buying record is obtained; a common word base is adopted, and in addition, common words in the comment information are extracted according to at least one classification factor; a classification factor information word base is adopted to obtain the classification factor information of the common words in one-to-one correspondence to each classification factor; the classification factor information is recorded into the personnel information. The invention discloses a system for obtaining personal information of the buyers, a method and a system for presuming the personnel information of the buyers, a method and a system for classifying commodities and a method and a system for retrieving the commodities. The resources of the personnel information of the buyers are diversified, and are not provided in a way of completely relying on users, in addition, the commodity classification mode is diversified, and faster and more precise retrieval results can be provided for the users according to different user groups.

The invention discloses a rapid image retrieval method, relating to a method which retrieves and displays digital images; the method solves the problem of complicated retrieval input. The method of the invention comprises the following steps: the process of creating and storing the comparison table between an image name assembly and a simplified name alphabet on an image processing device; the process of storing images on a storage medium; the process of establishing an image file chain table on the image processing device based on the images stored on the storage medium; the process of creating a hash table of the image simplified names on the image processing device based on the alphabet of the comparison table and the image file chain table; and the process of retrieving the images on the storage medium by simplified names and displaying the thumbnails of the retrieval result on the image processing device through the hash table and the image file chain table. The method of the invention enables the operator to input the simplified names of the images during the retrieval process, so as to avoid the image names expressed in complicated vocabulary; therefore, the method requires small input amount and is simple to operate, which occupies small storage space and has fast speed in the middle process; the method is suitable for hand-held image browsing devices.