Mediums: Machine Learning, and Coding
Theme: A machine learning-assisted tool that translates AI generated images into multiple director-specific cinematic shot suggestions and generations
Tools: Javascript, CSS, HTML, Gemini API
Solo Project | 1 month | 2026
CINEGRAM
Painpoint
AI filmmaking tools have lowered technical barriers to creating cinematic visuals. While this democratization enables more people to produce high-quality visual assets, many still lack the cinematic vocabulary for organizing them into coherent narratives.
Project Abstract
In response, this project proposes a system that goes beyond image generation by translating visual input into shot-based frameworks informed by established directors, helping users develop a deeper understanding of cinematic grammar and make more intentional storytelling decisions.
Function 1: Shot Type Classification
When users upload an AI-generated image, the system first identifies its shot type, such as close-up, medium shot, or full body, and recommends reference images with similar shot types from established directors.
Function 2: Cinematic Analysis
By selecting a reference, users can access a cinematic analysis across four aspects: composition, camera angle, lighting, and lens.
Function 3: Image Generation
Once a reference image is chosen, users can generate a new image that edits on the original input using the cinematic language of the selected reference image.
