Datasets:

SRCantona
/

panoramic_x-ray_tooth


	# Panoramic X-ray Tooth Dataset

	## Description
	This dataset contains over 500 panoramic dental X-ray images, accompanied by segmentation masks and annotation files. The dataset aims to support dental research, machine learning model development, and automated dental diagnostics. The images in this dataset represent a variety of dental conditions and tooth structures.

	## Dataset Structure
	The dataset includes the following components:

	1. X-ray Images: High-resolution panoramic X-ray images in JPG format.
	2. Machine Masks: Segmentation masks automatically generated by machine learning models, representing the detected teeth and surrounding structures.
	3. Human Masks: Segmentation masks manually annotated by dental professionals, serving as ground truth.
	4. JSON Annotation Files: Each X-ray image is accompanied by a JSON file containing polygon coordinates outlining the exterior boundaries of each tooth, along with class labels.

	### Example Files
	- 1.jpg: Example panoramic X-ray image.
	- Explanation: This image is a standard panoramic X-ray, capturing the full set of teeth, jaw, and surrounding bone structure in a single image. The process involves using a rotating arm of an X-ray machine that captures the dental arch. This type of imaging is widely used for dental diagnosis, as it provides a comprehensive view of the mouth.
	- Usage: This image can be used to detect dental anomalies, evaluate tooth alignment, and check for issues such as impacted teeth or bone loss.

	- 1.png: Corresponding machine or human mask.
	- Explanation: The mask highlights individual teeth by segmenting them from the X-ray image. In the case of machine-generated masks, computer vision models have been trained to detect and separate each tooth, while human masks are manually drawn for accuracy.
	- Usage: Overlay the mask on the X-ray to visualize segmentation. This is useful for training machine learning models to improve automatic segmentation accuracy.

	- 1.jpg.json: Annotation file containing labeled polygon data for each tooth.
	- Explanation: This JSON file contains detailed polygon coordinates that map out the boundaries of each tooth detected in the X-ray. Each entry in the file corresponds to a specific tooth number or region, providing crucial information for dental diagnostics and AI model training.
	- Usage: The polygon data can be used to train segmentation models, fine-tune existing algorithms, or serve as ground truth for evaluating model performance.

	## JSON File Breakdown
	Each JSON file includes:
	- Objects: A list of segmented teeth, each represented by a polygon.
	- ClassTitle: Labels representing tooth numbers.
	- Points: Polygon coordinates marking the exterior of each detected tooth.
	- Size: Width and height of the original image.

	Example Entry:
	```json
	{
	"classTitle": "8",
	"points": {
	"exterior": [[963, 587], [964, 569], [968, 546], ...],
	"interior": []
	}
	}
	```

	## Usage
	This dataset can be used to train segmentation models for tooth identification, dental health assessment, and automated diagnostics. The machine-generated masks can help in benchmarking, while human-annotated masks provide reliable ground truth for comparison.

	### Application Workflow
	1. Load the X-ray Images: Use the images for preprocessing and augmentation.
	2. Apply the Masks: Overlay machine or human masks to visualize tooth segmentation.
	3. Model Training: Utilize the masks and annotations to train segmentation models.
	4. Evaluation: Compare model output with human annotations for accuracy assessment.

	## Author and Attribution
	- Annotations by: GhazalehHITL
	- Dataset Curator: [Your Name or Profile Link]
	- Source: This dataset is publicly available for non-commercial research and development.

	## Download
	Click the link below to download the dataset:
	[Download Dataset](#)

	## License
	This dataset is licensed for educational and research purposes. Proper attribution is required for any publications or projects utilizing this dataset.

	---
	If you have any questions or require further information, feel free to reach out via [Your Contact Information].