Zoe Zhang
Media art, generative AI

Time Corridor

A modern corridor that slowly dissolves into fog, stone and caves, and opens into a shanshui landscape.

Film, 1 min 11 s. Press play, with sound on.

Type
AI-generated moving image
Duration
1 min 11 s
Built with
Stable Diffusion (Automatic1111), ComfyUI, Hailuo AI image-to-video
Context
Media Arts & Technology, UC Santa Barbara

Time Corridor is an experimental AI-driven film about the passage from built space into landscape. It begins in a contemporary corridor, glass and steel receding into perspective, and slowly lets the architecture dissolve into fog, stone and caves, until it opens into expansive mountains drawn from traditional Chinese shanshui painting.

The project asks how generative AI can reinterpret a spatial environment across different aesthetic and cultural contexts. Instead of producing a single image, it builds one continuous transition in which architecture becomes nature, and nature becomes a symbolic landscape: a corridor through time as much as through space.

Different tools were chosen for different stages of the journey. Stable Diffusion (Automatic1111) generated the early architectural scenes, and ComfyUI the later mountain landscapes, where the composition needed tighter control. The stills were then animated with Hailuo AI’s image-to-video system, which interpolates motion between a start frame and an end frame.

Process

  1. Architecture.Contemporary corridors generated in Stable Diffusion (Automatic1111).
  2. Landscape.Fog, stone, caves and shanshui mountains composed in ComfyUI.
  3. Motion.Each pair of stills joined into movement with Hailuo AI image-to-video.

Back to all work