Synthetic Data for Computer Vision

Naser Damer, Fraunhofer Institute for Computer Graphics Research IGD, Synthetic Biometrics: Why? What? and How?

Time-Resolved MNIST Dataset for Single-Photon Recognition

• TONO: a synthetic dataset for face image compliance to ISO/ICAO standard
• Improving Post-Earthquake Crack Detection using Semi-Synthetic Generated Images
• DiffAugment: Diffusion based Long-Tailed Visual Relationship Recognition
• Neural Transcoding Vision Transformers for EEG-to-fMRI Synthesis
• NeRFmentation: NeRF-based Augmentation for Monocular Depth Estimation
• NToP: NeRF-Powered Large-scale Dataset Generation for 2D and 3D Human Pose Estimation in Top-View Fisheye Images
• Training and Benchmarking Leukocyte Sub-types Classification Methods with Synthetic Images
• SurgicaL-CD: Generating Surgical Images via Unpaired Image Translation with Latent Consistency Diffusion Models
• Diffusion-based Synthetic Dataset Generation for Egocentric 3D Human Pose Estimation
• A CycleGAN Model to Synthesize Missing and Unpaired MRI Sequences for Under-Represented Multiple Sclerosis Lesions
• The Impact of Balancing Real and Synthetic Data on Accuracy and Fairness in Face Recognition
• DreamTexture: High-Fidelity Synthetic 3D Data Generation through Decoupled Geometry and Texture Synthesis
• Control+Shift: Generating Controllable Distribution Shifts
• Comparative Analysis of Synthetic and Real Melanoma Images in AI-Driven Diagnosis
• How Knowledge Distillation Mitigates the Synthetic Gap in Fair Face Recognition
• Synthetic Generation of Dermatoscopic Images with GAN and Closed-Form Factorization

DALDA: Data Augmentation Leveraging Diffusion Model and LLM with Adaptive Guidance Scaling

Contextual Knowledge Pursuit for Faithful Visual Synthesis