Design and Evaluation of Emotional Prompts for Generating Videos with High Emotion Consistency
摘要
Artificial Intelligence Generated Content (AIGC) is an advanced AI technology that can automatically create multimedia content including text, images, sound, and video. With the development of technologies such as ChatGPT, AIGC has made significant progress in natural language processing and exhibits excellent fluency and diversity in human-computer interaction. Text prompt generation is a new challenge for AIGC. Existing models can not only generate video content containing text and image elements, but also strive to weave a world of stories full of emotion and logical coherence to cater to users’ personalized desires. In the process of generating with text prompts, the clever use of adjectives is key to creating engaging scenarios. This study focuses on the design of Emotional Prompts using emotional words, adjectives, and onomatopoeic words to generate videos and sounds with corresponding emotions and further evaluate the emotional consistency of audio and video, enhancing the emotional depth of the content and resonating with the audience.