30results about How to "Intended" patented technology

The invention provides a method for acquiring search terms, a server and a method and a system for recommending the search terms. The method for acquiring the search terms includes setting a tag library; judging whether received application keywords are fussy keywords or not; matching the received application keywords with corresponding tags if the received application keywords are the fuzzy keywords; acquiring corresponding categories according tot the matched tags; summarizing the acquired categories and finding out the categories with the highest frequency of occurrence; finding out popular tags corresponding to the categories with the highest frequency of occurrence and acquiring the recommended search terms. The multiple tags, the multiple categories and a plurality of application keywords are stored in the tag library, each category corresponds to multiple tags, each application keyword corresponds to at least one tag, and each tag corresponds to at least one category. The method for acquiring the search terms, the server and the method and the system for recommending the search terms have the advantages that potential demands of users can be mined under the condition that subjective search purposes of the users are undefined, or demands of the users can be refined, so that search results can effectively conform to intention of the users, and the methods, the server and the system are high in practicality.

The hand region (Rh) of an operator is detected from a photographic image, the center (Po) of the palm of the hand and the center position (Wo) of the wrist are identified, the origin coordinates (Cho) of a coordinate system for the hand and the direction of a coordinate axis (Chu) of the coordinate system for the hand are set, and a feature amount (D15h) for the movement of the hand is calculated. Furthermore, the coordinate system for the hand is used to detect the shape of the hand in the hand region (Rh), and a feature amount (D14) is calculated for the shape of the hand. Furthermore, a feature amount (D15f) for the movement of a finger is calculated on the basis of the feature amount (D14) for the shape of the hand. A gesture is assessed on the basis of the calculated feature amounts. The assessment of a gesture takes into account differences in the direction in which a hand is moved, or differences in the angle of a hand placed in an operation region, so misrecognition of operations can be reduced.

The invention discloses a commodity picture searching method and device, and belongs to the technical field of searching. The method comprises the following steps that: extracting a plurality of firsttags from a picture input by a user, and obtaining the first tag selected by the user from the plurality of first tags; extracting a second tag from commodity description information input by the user; and according to the selected first tag and second tag, carrying out commodity picture searching, and outputting a commodity picture searching result. By use of the embodiment of the invention, thefirst tag and the second tag are combined to carry out the commodity picture searching so as to enable a search result to more conform to user intentions, so that the accuracy of the picture searching result is improved, the user can quickly and accurately search the corresponding commodity, and user experience is improved.

The present invention relates to a searching system for object-based images, which is capable of effectively searching for a registered image, video, or related information having a similar configuration of object-based information to that of information inputted by a user. The searching system for object-based images according to the present invention is to classify or cluster the registered images or videos on the basis of an object according to feature points or attributes of the feature points, and simply search for a matched registered image or video so as to provide a corresponding result or related information to a user terminal such as a PC, tablet, mobile terminal or connected TV, when a user uploads or searches for, by voice or text, an image or a video via the user terminal.

A display apparatus is provided. The display apparatus includes: an output unit; a voice collector which collects a user's voice; a first communication unit which transmits the user's voice to a first server and receives text information which corresponds to the user's voice; a second communication unit which transmits the received text information to a second server; and a controller which, when response information which corresponds to the text information is received, controls the output unit to output a system response which corresponds to an utterance intention of the user based on the response information, and when the user's utterance intention is related to at least one of performance of a function of the display apparatus and a search for a content, the system response includes an additional question which relates to the at least one of the performance of the function and the search for the content.

The present invention provides an input data processing method and apparatus. According to the method provided by embodiments of the present invention, at least one piece of input data of a user is acquired, and semantic data is acquired according to the at least one piece of input data, so that push information can be acquired according to the semantic data, and the push information is provided to a user. The push information is acquired not directly according to the input data of the user but according to the semantic data acquired according to the input data of the user, so that the push information provided to the user can be better conform to the intention of the user, thereby improving the flexibility of input data processing.

The invention discloses a hierarchical image segmentation method based on multi-scale edge cues, comprising the following steps: performing superpixelation on an input image to obtain an over-segmented region of the image; performing multi-scale content style separation on the input image, thereby extracting The content components of the image, and use the edge detection method to detect the edge information of the image content components, and obtain a multi-scale edge probability map; measure the similarity between the over-segmented regions according to the multi-scale edge clues, and perform iterative region merging to generate a hierarchy Merge tree: establish an energy function based on the hierarchical merge tree and prior information, and use the dynamic programming method to solve the energy function to optimize the hierarchical merge tree structure and realize the hierarchical segmentation of the image.

The embodiment of the present invention provides a voice input method, device and system, the method includes: when receiving the voice data sent by the client, judging the input environment type when the voice data is input; One or more candidate recognition texts corresponding to the voice data and matching the input environment type; sending the one or more candidate recognition texts to the client; wherein, the one or more candidate recognition texts for display on the client side. In the embodiment of the present invention, different input environment types have different language characteristics. Based on different input environment types, different language models are selected for recognition, which improves the performance of the language model and also improves the recognition accuracy of voice data, which is more in line with User intent improves user experience.

The invention discloses an imaging method and an imaging device. The imaging method includes the following steps of detecting first triggering operation, responding to the first triggering operation so as to obtain shooting guide information, detecting second triggering operation, and responding to the second triggering operation so as to shoot images.

The invention discloses a construction method and apparatus of three-dimensional simulation models and model bases. The construction method of three-dimensional simulation models and model bases includes the steps: acquiring a model base, classifying the models in the model base to form sub model bases; according to the semantic information, searching the model base to determine the import paths which are corresponding to the sub model bases; receiving different models in the import paths, and taking the different models as the content model and the style model; by means of a style content transfer method, synthesizing the content model and the style model into a new model, and storing the new model into a new model base, wherein the style content transfer method is to extract a bounding box of the style model as a bounding box of the content model; and synthesizing the new model base into the sub model bases to enrich the model bases until all the models in the sub model bases are executed. The construction method and apparatus of three-dimensional simulation model and model base have the advantages of effectively realizing variety of three-dimensional simulation models, enrichingthe model bases, effectively reducing the calculated amount of constructing a large quantity of three-dimensional model bases with rich contents and styles, and improving the construction efficiencyof the three-dimensional simulation models and the model bases.

The invention discloses a code search recommendation device and method based on open source knowledge. The device of the present invention includes a code segment generation module, a code feature extraction module, a text preprocessing module and a code search module. The steps of the method of the present invention are: generating a code fragment library; extracting code fragment text features; generating code fragment theme features; generating code fragment structural features; generating code fragment development features; Query similarity calculation; complete recommendation based on comprehensive score. The present invention uses open source knowledge to build a code feature measurement system, extracts code segment features from multiple angles, and uses similarity calculation and weighted calculation methods to obtain comprehensive scores of code segments and complete code search recommendations, perfecting the code measurement angle and improving code search recommendations accuracy.