5 results about "Retrieval result" patented technology

The invention discloses an image retrieval method comprising the following steps of: firstly, carrying out grey level transformation and noise reduction treatment on a scene image acquired by an image module to obtain a scene grey image; then, detecting a target object outline in the scene grey image; and finally, updating the target grey image or the feature description of the target grey image in the target object outline to an image retrieval server end for image retrieval. Correspondingly, the invention also discloses an image retrieval device, a mobile terminal and an image retrieval system. Before image retrieval, a target object is extracted precisely through outline detection, so that the influence of an image background to an image retrieval result is effectively reduced, and the retrieval precision is improved.

The invention provides a multilayer retrieval method based on word segmentation. The method comprises the steps of obtaining retrieval basic split words of a retrieval text; combining the retrieval basic split words to form retrieval combined split words of the retrieval text; using the retrieval basic split words and the retrieval combined split words of the retrieval text to search for text datain a corpus to obtain a retrieval result of the retrieval text.

The invention relates to a relevance feedback method based on a transfinite learning machine. The relevance feedback method comprises the following steps: inputting an inquiry image; retrieving the image to obtain a retrieval result, and marking a result by a user; respectively extracting an SIFT characteristic, a Colour characteristic and an LBP characteristic from the marked image; training three basis classifiers by utilizing the three kinds of characteristics; respectively putting the image in a retrieval image library in the three basis classifiers, voting according to a prediction result, and automatically marking each unmarked image; training and updating the classifiers again; classifying the image in the image library; and returning a result. The relevance feedback method disclosed by the invention is established on the basis of the transfinite learning machine; human-computer interaction is carried out by introducing the inquiry intention of human beings; learning data are enriched by effectively utilizing the unmarked image in the image library; therefore, the image feedback precision is greatly increased; furthermore, the processing speed is controlled well; expression of the image in a computer accords with understanding of human beings to image semantics well; and thus, the relevance feedback method has a good feedback effect.

The invention discloses an online translation model selection method for statistical machine translation, which includes two stages of training and translation. The training stage includes: collecting bilingual parallel corpora, and dividing the bilingual parallel corpus into different sub-corpora according to types; The corpus trains the candidate translation model; builds an index for the sub-corpus, and obtains the corpus index file; the translation stage includes: inputting the text to be translated, and retrieving sentences similar to the sentences in the text to be translated from the corpus index file; The candidate translation model corresponding to the sub-corpus where the sentence is located, selects the final translation model from all the candidate translation models; translates the input text to be translated according to the final translation model, and obtains the final translation result. The invention well solves the problem that the statistical machine translation system cannot adapt to input texts in different fields, and can effectively improve the translation quality of the statistical machine translation system.