Dataset Structure:
- Classes: 144 (12 hours x 12 minutes)
- Subsets: Train, test, validation
- Image Specifications: 224x224x3 jpg format, randomly rotated (0, -90, +90, 180 degrees)
- Hands: Rectangular, black or in 5 different colors on darker faces, minute hand at exact positions (0, 5, 10, etc.), hour hand adjusted for realism.
Additional Details:
- The dataset’s design aims to provide a realistic analog clock appearance, including proper hour hand positioning.
- Models trained on this dataset might not perform well on real-world images due to fixed design constraints in the synthetic images.
Use Cases: This dataset is ideal for training image classification models to recognize specific times on clock faces, offering a structured approach for developing algorithms in time-based image recognition tasks.