FLAME-2: Thermal Infrared Fire Detection Autoencoder Weights

FLAME-2 thermal infrared fire detection dataset includes autoencoder weights and YOLO pseudo-labels. The dataset appears to be hosted on Kaggle. Its specific size, creation date, and author are unknown.

Use Cases

• Train or fine-tune autoencoder models for anomaly detection in thermal imagery based on the provided weights.
• Develop fire detection systems using thermal infrared data based on the dataset's focus.
• Generate pseudo-labels for training object detection models like YOLO based on the included pseudo-labels.
• Benchmark the performance of different model architectures on thermal fire detection tasks.

Strengths

• Includes pre-trained autoencoder weights, which can reduce training time and computational cost.
• Provides YOLO pseudo-labels, which can serve as a starting point for supervised object detection tasks.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Kaggle