Doğal dil işleme temelli görsel senaryo taslağı motoru geliştirilmesi

Storyboards are essential to ensure that all visual elements appear in a compatible way according to the scenario. At this point, it should be discussed what might be a visual element. By choosing the easiest way, all objects can count as a visual element; but their facades may change due to the time interval of the main event. Storyboards are generally created using texted scenarios. To extract an object and their actions, Natural Language Processing Techniques can be used. Even if this information from the scenario is obtained, storyboard creation is a laborious process that requires technical knowledge. For this reason, storyboard creators must know the drawing tools well. Knowledge extraction is a hard process. In this dissertation, the main knowledge that should be extracted is roles, which are known as thematic roles. On the other hand, word placements should be examined in accordance with their places in the sentence. This analysis is done using grammatical roles. Thus, in this thesis, Natural Language Processing is used to find thematic and grammatical roles. In this study, it is aimed that storyboard creators will direct their talents to creativity rather than using these tools through a new tool that will be developed. Development progress of the tool contains two main stages which are Natural Language Processing and drawing of the extracted objects and its actions/motions on a storyboard.