Table 3 Dataset details before and after preprocessing, augmentation, and balancing.
From: Real time fire and smoke detection using vision transformers and spatiotemporal learning
Dataset | Stage | Fire count | Non-fire count | Total count |
|---|---|---|---|---|
NASA space apps challenge | Before preprocessing | 755 | 244 | 999 |
After preprocessing (Resizing + Normalization) | 755 | 244 | 999 | |
After balancing (Targeted augmentation) | 755 | 755 | 1,510 | |
Kaggle fire videos | Before preprocessing | 12,000 | 4,800 | 16,800 |
After preprocessing (Frame extraction + Normalization) | 12,000 | 4,800 | 16,800 | |
After balancing (Targeted augmentation) | 12,000 | 12,000 | 24,000 | |
Combined dataset | – | 12,755 | 12,755 | 25,510 |
