Description
|
In order to facilitate a new object detection and image enhancement research particularly in the low-light environment, we introduce the Exclusively Dark (ExDark) dataset (CVIU2019). The Exclusively Dark (ExDARK) dataset, which to the best of our knowledge, is the largest collection of low-light images from very low-light environments to twilight (i.e 10 different conditions) with 12 object classes (similar to PASCAL VOC) annotated on both image class level and local object bounding boxes.
|
Notes
| The images are kept in individual class folders following the image class labels: Bicycle - 652 images Boat - 679 images Bottle - 547 images Bus - 527 images Car - 638 images Cat - 735 images Chair - 648 images Cup - 519 images Dog - 801 images Motorbike - 503 images People - 609 images Table - 505 images Total : 7,363 images For the experiments in our paper that involves training, the data are split as follows: (a) Training - 3,000 images (250 images per class) (b) Validation - 1,800 images (150 images per class) (c) Testing - 2,563 images |