Floor Play

Data Preparation & Training

Data Source: SESYD

SESYD "Systems Evaluation SYnthetic Documents" is a database of synthetical documents, with the corresponding groundtruth, produced using the 3gT system.

Here, the floor plan collection is used.

Model: YOLOv5

YOLO is a family of object detection architectures and models pretrained on the COCO dataset, and represents Ultralytics open-source research into future vision AI methods, incorporating lessons learned and best practices evolved over thousands of hours of research and development.

Process

https://colab.research.google.com/drive/1-zGjeJG--licpHJ3KR1uwHx4uZYQiWR7?usp=sharing

Inference

https://colab.research.google.com/drive/1C6vpt7aclephCBg7TMBe0uSH1LuH_d7q?usp=sharing

Limitation

Because in training data, the windows are like

window1

and

window2

Doors are like

door

if your input image uses different symbols, the model can't recognize them.