Article Preview
TopIntroduction
The short video industry has developed rapidly in China and has become increasingly mature. With the development of modern We Media and the emergence of We Media short videos, it has also had a certain impact on international Chinese language education and teaching. Video entertainment is becoming more and more popular in China today. Its language expression and cultural factors are very beneficial for foreign students to learn Chinese (Tian& Hu, 2021). Movies, documentaries and other film and television works have appeared more and more in Teaching Chinese as a Foreign Language classes. For foreign students, film and television works give consideration to language training in listening, reading (subtitles), speaking, and other aspects. Especially in the teaching of Chinese culture as a foreign language, film and television works can make them feel the abstract Chinese culture more intuitively and the cultural habits hidden behind the social image (Chen et al., 2020). There is no doubt that audiovisual media has great potential as an auxiliary means of teaching, but there are still many problems in its application in the cultural classroom, such as the inappropriate selection of film and television works, the long-playing time, and the lack of timeliness of content.
Multimedia-assisted teaching of Chinese as a foreign language is the general trend. Nowadays, the teaching of Chinese as a foreign language shows a trend of diversification. More and more teachers of Chinese as a foreign language are trying new teaching methods to help foreign students learn more language and cultural knowledge in a relaxed and pleasant teaching environment (Qian et al., 2021). In this situation, online short video has gradually entered the public’s view with its unique interest, and it is also expected to enter the classroom of teaching Chinese as a foreign language (Cao et al., 2019). Based on the above two points, in this paper the author preliminarily discusses how to use the entertainment form of short online video as an auxiliary means of cultural teaching, and tries to integrate this emerging thing into the cultural teaching of Chinese as a foreign language, bringing vitality and vitality to the teaching of Chinese as a foreign language. What is the relationship between language teaching and culture teaching? How should the content of culture teaching be defined? These are the theoretical basis for the short video subtitle generation algorithm to integrate into international Chinese language education and cultural teaching. Traditional subtitles need manual editing, which is inefficient. For many languages, subtitle editing needs to master many languages, which is unrealistic for most people. In addition, after manually editing subtitles, it is also necessary to align the voice and subtitles in time, which requires the editors to determine the start time and end time of the audio corresponding to the subtitles (Singh et al., 2022). In addition, the short video service may require various subtitle effects, and it is difficult to add these animations for manual editing. In addition, most video editing software on the market can only add animation after opening members, which is costly.
Video caption generation technology mainly includes two important challenges, namely video understanding and language generation. First of all, video understanding not only includes the description of the scene in the current video, but also needs to explore its deep logic. Secondly, on the premise that the generated content is correct, it is also necessary to consider the accuracy of the language sequence before and after. How to take these two aspects into account is the key to the current video caption generation method. The current video caption generation methods can be summarized into the following two types: Time-series based method and multimodal method (Li et al., 2019). The problem of video caption generation the author investigated in this research is difficult to study, mainly because it is a cross modal task, which requires sufficient understanding of the existence of different modes, and how to transform each mode. In the process of mode conversion, the inevitable problem is that there is strong heterogeneity between the two modes, which makes the generated statements not real enough. Secondly, it is also necessary to analyze the objects in the video, as well as the relationship between each object, and extract key information from them to generate statement annotation.