Dataset Versions
v1
2024-10-11 5:25am
Generated on Oct 11, 2024
736 Total Images
View All Images
Dataset Split
Train Set 96%
708Images
Valid Set 3%
20Images
Test Set 1%
8Images
Preprocessing
Auto-Orient: Applied
Isolate Objects: Applied
Dynamic Crop: Class: phone_back_segment
Resize: Stretch to 640x640
Auto-Adjust Contrast: Using Adaptive Equalization
Tile: 2 rows x 2 columns
Augmentations
Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Rotation: Between -15° and +15°
Hue: Between -21° and +21°
Brightness: Between -23% and +23%
Exposure: Between -10% and +10%
Blur: Up to 2.5px
Noise: Up to 1.84% of pixels
Cutout: 5 boxes with 10% size each
Bounding Box: Flip: Horizontal
Bounding Box: 90° Rotate: Clockwise, Counter-Clockwise