horror-object-detector

Horror Object Detector

Custom YOLO11n object detection model for the AI Horror Game.

Classes

The model detects 8 object classes:

• chair
• bottle
• table
• book
• box
• lamp
• door
• window

Training

This model was fine-tuned from pretrained YOLO11n weights using a custom Open Images based dataset.

Input

An image containing game-environment objects.

Output

The model detects objects and their bounding boxes. The game API converts the detections into object counts.

Example:

{
  "objects": {
    "chair": 4,
    "bottle": 2,
    "table": 1,
    "book": 0,
    "box": 1,
    "lamp": 2,
    "door": 1,
    "window": 3
  }
}