Expression input method and device based on semantic understanding

Abstract

The invention discloses an expression input method and device based on semantic understanding, and relates to the technical field of input methods. The method comprises the following steps: obtaining text contents corresponding to an input sequence; performing word segmentation on the text contents, and extracting features of each text based on a result of the word segmentation; using features of each text to build input vectors, and substituting an emotion classification model to classify so as to determine emotion labels of the text contents; respectively obtaining expressions of each theme under the emotion labels based on the correspondence between the emotion labels and the expressions of each theme; sorting the expressions of each theme, and using the sorting result as a candidate item to be showed on a client side. According to the method and device provided by the invention, semantic recognition can be realized aiming at the text contents corresponding to the input sequence of a user to obtain the emotion labels which are the best match of semantics of the text contents, and then the expressions under each corresponding theme are obtained based on the emotion labels and are used as the candidate item to be sent to the user, so that the user can conveniently input the expressions, the expression input efficiency is improved, and a rich and wide-range expression resource is provided for the user.

Description

[0001] The embodiment of the present invention claims the priority of the Chinese patent application submitted to the China Patent Office on June 6, 2014, the application number is 201410251399.0, and the title of the invention is "A Method and Device for Expression Input Based on Semantic Understanding", the entire content of which is passed References are incorporated in the embodiments of the present invention. technical field [0002] The invention relates to the technical field of input methods, in particular to an expression input method and device based on semantic understanding. Background technique [0003] The input method is an encoding method used to input various symbols into a computer or other equipment (such as a mobile phone). Common input methods include Sogou input method, Microsoft input method and so on. [0004] Traditional emoticon input generally has several situations: one is that the platform itself has an emoticon input module, such as the emotico...

AI Technical Summary

Problems solved by technology

Then, within a limited period of time, the time and cost it takes to find suitable emoticons one by one in many emoticon databases will become a bottleneck for users to effectively use various emoticons to enhance communication.

[0009] 2. Considering the operating cost of using emoticons for users, the emoticon package producers will also simplify the content of emoticons as appropriate, which also restricts the development and widespread use of chat emoticons to some extent

[0010] 3. Most chat tools only provide default emoticons

The default emoticons are relatively monotonous, and more abundant and diversified themed chat emoticon resources can effectively improve the favorability of chatting with friends. However, in order to use these emoticons, users need to go through many online operation steps, obtain emoticon package information from various channels and Download the emoticon pack to the local, and sometimes it needs to be manually loaded to use the emoticon pack normally

For users who are not familiar with operations or who are not patient enough, the time and cost it takes to successfully obtain and install a suitable emoticon package from network resources may cause them to choose to give up

[0011] 4. For the downloaded emoticons, if the user switches the input scene such as the chat platform, the emoticons need to be downloaded or updated again, and the user's commonly used emoticon collection information also faces the problem of transplantation

If it is not specially organized, many multimedia resources such as photos of exaggerated expressions of celebrities and political figures, GIFs, etc. cannot be used as candidate expressions in a timely manner. The input content is too single, and it cannot quickly and accurately match the user's current input, which reduces the user's input efficiency.

Images