The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
This is a modification of the classic MNIST dataset, transforming it into a dataset tailored for object detection tasks. This modified dataset includes images of hand-written digits, similar to the original MNIST dataset.
However, unlike traditional digit classification, each image is now equipped with bounding boxes, precisely defining the area where a given digit is located. This has allowed me to create a dataset that is well-suited for solving object detection problems.
Key information about the dataset:
Number of examples: 60000 trains images, 10000 test images Image size: 28x28 pixels number of bounding boxes per image: 1, and label from 0-9 digit
This dataset has immense potential for applications in the field of machine learning, particularly in tasks related to object detection. You can use it for training and evaluating models that combine classification and regression tasks. Solutions based on this dataset can find applications in areas such as automatic identification of digits in images or more advanced detection tasks.
import numpy as np
# Loading data from the 'mnist_object.npz' file
data = np.load('/kaggle/input/mnist-for-object-detection/mnist_object.npz')
# Reading variables containing the data
X_train, y_train, y_train_bboxes, X_test, y_test, y_test_bboxes = data['train_images'], data['train_labels'], data['train_bboxes'], data['test_images'], data['test_labels'], data['test_bboxes']
- Downloads last month
- 5