language	license	pretty_name	tags	annotations_creators	task_categories	dataset_info	configs
en	apache-2.0	Basic Shapes Object Detection	object-detection simple example basic-geometric-shapes	machine-generated	object-detection	features name dtype image_id int64 name dtype image image name dtype width int32 name dtype height int32 name sequence objects name dtype id int64 name dtype area int64 name sequence length bbox float32 4 name dtype category class_label names 0 1 2 Square Circle Triangle	config_name data_files default split path train data/train-* split path test data/test-*

Basic Shapes Object Detection

Description

This Basic Shapes Object Detection dataset has been created to test fine-tuning of object detection models. Fine-tuning some model to detect the basic shapes should be rather easy: just a bit of training should be enough to get the model to do correct object detection quite fast.

Each entry in the dataset has a RGB PNG image with a white background and 3 basic geometric shapes:

• A blue square
• A red circle
• A green triangle

All images have the same size. Each image has exactly 1 square, 1 circle and 1 triangle, with their fixed colors. Each entry in the dataset has consequently 3 bounding boxes. The shapes do not overlap.The category IDs are 0, 1 and 2, corresponding to the labels Square, Circle and Triangle.

The dataset has exactly the same structure as the https://huggingface.co/datasets/cppe-5 dataset, but fine-tuning some model to this dataset with basic geometric shapes should require considerable less training compared to the cppe-5 dataset. Once you have tested your fine-tuning code on this dataset, it should also work on more complicated datasets such as the cppe-5 dataset.

Links

The Python code to generate the images can be found at https://github.com/DriesVerachtert/basic_shapes_object_detection_dataset The dataset can be downloaded from https://huggingface.co/datasets/driesverachtert/basic_shapes_object_detection

Structure

The bounding boxes are in COCO format (x_min, y_min, width, height).

License

This dataset is released under Apache 2.0.

Usage

from datasets import load_dataset
dataset = load_dataset("driesverachtert/basic_shapes_object_detection")