🌳 HNIVision — Five vision methods for Human-Nature Interaction

A unified pipeline running 5 CV methods on the same image → merged into a 4-layer HNI schema (human · nature · activity · meaning).

  • Click a gallery example → instant pre-computed results
  • Upload your own image → live inference (~60-120s on CPU)

🔗 GitHub · ResNet weights
Pre-computed gallery (click for instant)

Backends: 🏷️ Google Vision · 📊 ResNet-50 · 📦 YOLO26-s · 🎨 SAM2+SegFormer · 🤖 GPT-4.1-mini