3D Model Training for Anatomical Landmark Prediction
Budget: $250 – $750 USD
I need an expert to train a 3D model for predicting anatomical landmarks in CTA DICOM volumes. Each volume has 20 annotated landmarks, and the model must achieve prediction accuracy within 1.5 mm.
Requirements:
- Data preprocessing: normalization, data augmentation, noise reduction
- Architecture: U-Net
- Label representation: heatmaps
- Custom loss functions and robust training strategies
- Comprehensive evaluation metrics
Ideal Skills and Experience:
- Strong background in 3D deep learning
- Proficiency with U-Net and heatmap generation
- Expertise in medical imaging and DICOM format
- Experience in rigorous model evaluation and error analysis
Your work will directly contribute to advancing precision in medical imaging tasks. Please provide a portfolio showcasing relevant projects.
Requirements:
- Data preprocessing: normalization, data augmentation, noise reduction
- Architecture: U-Net
- Label representation: heatmaps
- Custom loss functions and robust training strategies
- Comprehensive evaluation metrics
Ideal Skills and Experience:
- Strong background in 3D deep learning
- Proficiency with U-Net and heatmap generation
- Expertise in medical imaging and DICOM format
- Experience in rigorous model evaluation and error analysis
Your work will directly contribute to advancing precision in medical imaging tasks. Please provide a portfolio showcasing relevant projects.