Scene Descriptions for 12 Film Scenes Used in Generative AI Video Analysis

Twelve Korean-language scene descriptions document the visual conditions of film scenes from four cinematic styles: German Expressionism, Soviet Montage, Italian Neorealism, and the French New Wave. The descriptions, created by Jinyu Kim, detail spatial composition, characters, objects, lighting, color, camera work, and editing. These texts served as intermediate research documents for generating controlled English prompts for AI video models Kling 3.0, Vidu Q3, and PixVerse V6.

Use Cases

• Analyzing the translation of visual film features into text prompts based on documented elements like spatial composition and lighting.
• Comparing scene descriptions across different film styles (German Expressionism, Soviet Montage, etc.) for stylistic analysis.
• Training or evaluating text-to-video models using structured, style-specific scene descriptions.
• Studying the role of intermediate multilingual documents (Korean to English) in generative AI workflows.

Strengths

• Focuses on 12 film scenes from four distinct and historically significant cinematic movements.
• Descriptions document multiple specific visual conditions, including camera position, movement, editing, and style-specific features.
• Data is structured into 12 separate text files, each corresponding to a specific scene code.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is small in scale at 160.5 KB, covering only 12 scenes.

Provenance

Source

Jinyu Kim via figshare

Collection Method

Likely manually created as intermediate research documents for a generative AI analysis project.

Freshness

Last updated 2026-06-04 20:57:23; freshness should be verified.