Cross-media star emotion accompanying interaction system based on machine learning

Abstract

The invention discloses a cross-media star emotion accompanying interaction system based on machine learning. The system comprises: an input content recognition module, which is used for recognizing input content in various media forms, and reading the meaning of the input content; an interactive content matching module, which is used for matching a verbal skill file corresponding to the meaning of the input content; a star real person image module, which comprises a sound ray model and a virtual image model, and is used for simulating sound ray characteristics and appearance skin of a star real person; and an output module, which is used for outputting the verbal skill file content matched with the input content. The technical scheme is adopted to define that the interaction system has star personal attributes, stimulates the sound ray characteristics and appearance skin of a star real person and determines the verbal skill characteristics of the star real person according to the defined star personal attributes of the interaction system, and response content for different input content is formed through machine learning, so that a user can obtain relatively real experience of interacting with the star.

Description

technical field [0001] The invention relates to a machine-learning-based cross-media star emotional companion interaction system, and belongs to the technical field of designing voice robots. Background technique [0002] Nowadays, there are many intelligent voice devices or robots, such as Cortana, Siri, OkGoogle in smartphones, Google Home, Amazon Echo in personal assistants, interactive voice response systems, telephone robots, customer service robots, and telemarketing robots in voice robots. Their working principles are roughly the same. A complete voice interaction system usually has three typical modules. Automatic Speech Recognition (ASR) converts sound into text, and Natural Language Processing (NLP) interprets the meaning of text. Come out, give feedback, and finally convert the output information into sound through speech synthesis (Text to Speech, TTS). [0003] Among the human-computer interaction methods, voice interaction is the most natural way of interactio...

AI Technical Summary

Problems solved by technology

In addition, in the current human-computer interaction system, there are still the following defects: 1. The interactive media is single, mostly a single media interaction of text or voice; 2. It is difficult to customize, and it is impossible to set different topic content and Different characteristics; 3. The deployment platform is single, which can only be deployed on the cloud platform, and the operation and maintenance costs are high; 4. The interaction mode is single, and it can only passively wait for user access, and cannot actively communicate; 5. The interaction scene is single, and cannot be based on the real-time environment and User portraits give different answers

Images