VisionDepth3D

VisionDepth3D
The All-in-One AI-Powered 3D Suite for Creators

VisionDepth3D brings cinematic-quality 3D conversion to your desktop using real-time AI depth estimation, stereo rendering, and GPU-accelerated post-processing — all wrapped in a clean, intuitive GUI.

🚀 Features:

• AI Depth Estimation — 20+ Hugging Face models (ZoeDepth, MiDaS, DPT, etc.)

• Stereo 3D Renderer — Half-SBS, Full-SBS, Interlaced, Anaglyph, VR180

• RIFE Frame Interpolation — ONNX runtime with 2x–8x FPS boosting

• Real-ESRGAN Super Resolution — 4x upscaling with CUDA-accelerated ONNX

• Smart Parallax Control — Separate depth sliders for foreground, midground, background

• Floating Window & Masking — Theater-style masking with smooth easing

• Preview & Testing — Live depth/3D previews in heatmap, SBS, interlaced formats

• 3D Player (VDPlayer) — Instantly preview results in Half-SBS/VR

• Audio Tools — Auto-attach audio from source using FFmpeg

💻 Requirements:

• Python 3.12

• CUDA (11.8+ recommended for GPU features)

• FFmpeg installed (included or separate)

📦 Installation:
Download the installer, extract, and run. Includes all dependencies and models.

❤️ Free & Open Source
Built by a solo developer. Donations go toward testing hardware, model research, and new features.