Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

This is a indoor building element dataset with 37 classes.

This dataset consists of 900 original images collected from the ETH building videos collected by Aria glasses, Excel sheet: 230116_Ernest-Bloch_Inventaire and web-scraped samples from gruner-reuse.ch using the script parse_grunerweb.ipynb. The dataset is randomly split into 720 training, 90 validation, and 90 test images.

After applying data augmentations, the final dataset contains a total of 2,340 images, including:

  • 2,160 augmented training samples
  • 90 validation samples
  • 90 test samples

Preprocessing

  • Auto-Orient: Applied
  • Resize: Stretched to 640Γ—640
  • Auto-Adjust Contrast: Using Adaptive Equalization

Augmentations

Each training example is augmented to generate 3 output variations with the following transformations:

  • Hue: Between βˆ’15Β° and +15Β°
  • Saturation: Between βˆ’25% and +25%
  • Brightness: Between βˆ’15% and +15%
  • Blur: Up to 1px
  • Noise: Up to 0.1% of pixels

Coco format:

 ```
data_new_coco/
β”œβ”€β”€ train/
β”‚   β”œβ”€β”€ images/
β”‚   └── annotations/
β”œβ”€β”€ valid/
β”‚   β”œβ”€β”€ images/
β”‚   └── annotations/
β”œβ”€β”€ test/
β”‚   β”œβ”€β”€ images/
β”‚   └── annotations/
β”œβ”€β”€ README.dataset.txt
└── README.roboflow.txt 
``` 

Yolo format:

 ```
data_new_yolo/
β”œβ”€β”€ train/
β”‚   β”œβ”€β”€ images/
β”‚   └── labels/
β”œβ”€β”€ valid/
β”‚   β”œβ”€β”€ images/
β”‚   └── labels/
β”œβ”€β”€ test/
β”‚   β”œβ”€β”€ images/
β”‚   └── labels/
β”œβ”€β”€ data.yaml
β”œβ”€β”€ README.dataset.txt
└── README.roboflow.txt
``` 
Downloads last month
28