ExDark Enhanced Dataset for Low-Light Object Detection

An enhanced version of the ExDark dataset, formatted for YOLO object detection models. The dataset contains images and corresponding labels for training and evaluation. The original ExDark dataset was created for research in low-light computer vision.

Use Cases

• Train YOLO models to detect objects in low-light conditions using the provided bounding box labels.
• Benchmark object detection performance on challenging, real-world low-light imagery.
• Fine-tune pre-trained detectors on the enhanced labels to improve robustness in dark environments.

Strengths

• Dataset is specifically enhanced for YOLO training, providing ready-to-use label files and configuration.
• Focuses on the challenging domain of low-light and nighttime object detection.

Limitations

• The exact number of images, classes, and annotation quality of the enhanced version is unknown.
• Potential for class imbalance or geographic bias inherited from the original ExDark dataset.

Provenance

Source

Enhanced version of the ExDark (Exclusively Dark) dataset.

Collection Method

Images were likely collected from various sources depicting low-light scenes and annotated with bounding boxes.