Facial Profile Soft Tissue Annotations for Orthodontic Diagnosis and Classification

Description

This dataset consists of 400 annotated facial soft tissue profile images curated for orthodontic research and deep learning model development. The images were annotated using the WebCeph tool, focusing on soft tissue landmarks such as the lips, nose, chin, and jawline. Additionally, the annotations include detailed skeletal and dental measurements, all validated by orthodontic specialists to ensure clinical relevance and precision. Each CSV file includes comprehensive patient data with the following fields: patients: Patient identifier (anonymized). skeletal_status: General skeletal classification. skeletal_maxilla: Skeletal condition of the maxilla. skeletal_mandible: Skeletal condition of the mandible. dental_status: Overall dental status of the patient. dental_overjet: Measurement of the overjet. dental_overbite: Measurement of the overbite. dental_upper_incisor_inclination: Inclination of the upper incisors. dental_lower_incisor_inclination: Inclination of the lower incisors. dental_interincisal_angle: Angle between the upper and lower incisors. dental_upper_incisal_display: Display of the upper incisors. soft-tissue_upper_lip: Soft tissue condition of the upper lip. soft-tissue_lower_lip: Soft tissue condition of the lower lip. The dataset provides five classification categories for the soft tissue profiles: Plane, Concave, Convex, Convex-Concave, and Concave-Convex. This classification assists in understanding soft tissue structure and planning appropriate orthodontic treatments. The CSV files contain detailed annotations for each patient, enabling research into the relationships between skeletal, dental, and soft tissue characteristics. This dataset is well-suited for developing deep learning models, such as convolutional neural networks (CNNs), to improve the accuracy and efficiency of orthodontic diagnosis and treatment planning. Key Features: Images: 400 high-quality profile images of patients’ facial soft tissues. Annotations: Comprehensive skeletal, dental, and soft tissue annotations provided in CSV format, using the WebCeph tool. Classes: Five classification categories for soft tissue profiles: Plane Concave Convex Convex-Concave Concave-Convex Format: Images in JPEG/PNG format, annotations in CSV format. Validation: All annotations validated by orthodontic specialists for clinical accuracy. Application: Orthodontic research, deep learning model development, AI-based diagnostics. Use Case: Ideal for training machine learning models for enhanced diagnosis and treatment planning in orthodontics, focusing on the relationships between skeletal, dental, and soft tissue profiles. Intended Audience: This dataset is designed for researchers, data scientists, and developers working in orthodontics, artificial intelligence, and medical imaging. It offers a valuable resource for advancing research in AI-powered diagnostic tools and soft tissue analysis in orthodontics.

Steps to reproduce

Access the Dataset Download the dataset from the provided link on Mendeley Data. The dataset includes 400 images of facial soft tissue profiles in JPEG/PNG format and corresponding annotation files in CSV format. Tools and Software Required WebCeph: A web-based tool for soft tissue analysis. (Optional, if users want to perform new annotations.) Deep Learning Framework: Install a framework such as TensorFlow or PyTorch for training the deep learning model. Python Libraries: Ensure that the following Python libraries are installed: OpenCV (for image processing) Pandas (for handling CSV annotations) NumPy Scikit-learn (for data preprocessing) Matplotlib (for visualizing results) Preprocessing the Data Load the CSV annotation files using Pandas to extract key landmarks for soft tissue analysis (e.g., lips, nose, chin). Normalize or scale the landmarks as necessary for deep learning model input. Optionally, augment the image data (e.g., flipping, rotation) to expand the dataset for better model generalization. Model Training Use a Convolutional Neural Network (CNN) architecture to classify the soft tissue profiles into five categories: Plane Concave Convex Convex-Concave Concave-Convex Input each image and its corresponding annotations into the CNN for training. Set aside a validation dataset to monitor the model’s performance during training. Evaluation and Testing Once the model is trained, evaluate its performance using metrics such as accuracy, precision, recall, and F1-Score on the test set. Compare the classification results with the annotated classes to assess the model's ability to classify the soft tissue profiles correctly. Optional Modifications Users may modify the CNN architecture, learning rate, or training epochs to experiment with different model performances. Additional data augmentation or normalization techniques can be applied to see how they affect classification accuracy.

Institutions

Categories

Funding

Licence