ThIRED is a first collection of thermal infrared images featuring pixel-level edge annotations. It is designed for computer vision tasks involving infrared imagery and edge detection.
Use Cases
- Train an edge detection model on thermal infrared images using the provided pixel-level edge annotations.
- Benchmark segmentation algorithms for thermal imagery based on the annotated edge data.
- Analyze the characteristics of edges in thermal infrared images compared to visible spectrum images.
Strengths
- First collection of its kind for thermal infrared images with edge annotations.
- Focuses on a specialized modality (thermal infrared) for computer vision.
Limitations
- Dataset size, row count, and specific image dimensions are unknown.
- Potential for limited geographic or scene diversity due to unspecified collection scope.
Provenance
- Source
- null
- Collection Method
- null
- Time Range
- null
- Freshness
- null
- Geography
- null