Important:
\nCombining these datasets for pretraining, augmentation, and multimodal analysis can help boost the accuracy of your aneurysm detection models.
", "messages": [ { "id": 3265360, "postDate": "2025-08-07T11:42:36.163Z", "content": "Important:
\nCombining these datasets for pretraining, augmentation, and multimodal analysis can help boost the accuracy of your aneurysm detection models.
", "rawMarkdown": "## 1. Intracranial Aneurysm CTA Images and 3D Models Dataset\n\n* **Description:** High-resolution CTA scans from 99 patients (105 aneurysm cases) plus 44 healthy controls. Includes 3D mesh models of aneurysms and parent vessels, as well as CFD-derived morphological and hemodynamic parameters.\n* **Usage Example:** Pretrain a risk-prediction model by reconstructing CTA into 3D and using CFD features.\n\n## 2. Time-of-Flight MRA of Intracranial Aneurysms with Interval Surveillance Imaging\n\n* **Description:** TOF-MRA scans from 63 patients (24 with longitudinal follow-up). Comes with vessel and aneurysm segmentation masks, clinical annotations, 3D models, and a 3D Slicer project.\n* **Usage Example:** Time-series analysis for aneurysm growth prediction or additional training data for segmentation models.\n\n## 3. Royal Brisbane TOF-MRA Intracranial Aneurysm Database\n\n* **Description:** TOF-MRA scans of 63 patients (85 aneurysm cases) with voxel-level segmentations and expert annotations.\n* **Usage Example:** Fine-grained segmentation training using 3D CNNs or weakly supervised detection.\n\n## 4. Lausanne UIA TOF-MRA Dataset (Di Noto et al., 2022)\n\n* **Description:** 284 subjects (157 aneurysm cases, 127 healthy controls). Includes 246 coarse (spherical) labels and 38 detailed labels—ideal for mixed weak/strong supervision.\n* **Usage Example:** Benchmarking weakly supervised learning frameworks.\n\n## 5. ADAM Dataset for UIA Segmentation (Timmins et al.)\n\n* **Description:** 113 subjects (93 unruptured intracranial aneurysm cases, 20 healthy). Provides voxel-level segmentation masks focused on central cerebral vessels.\n* **Usage Example:** External validation set for transfer-learning and generalization studies.\n\n## 6. IntrA: 3D Intracranial Aneurysm Dataset for Deep Learning\n\n* **Description:** Point-cloud and mesh representations of intracranial aneurysms with classification and part-segmentation benchmarks.\n* **Usage Example:** Training point-cloud networks or mesh-CNNs to learn shape features.\n\n## 7. Aneumo: Large-Scale Multimodal Aneurysm CFD Dataset\n\n* **Description:** 427 real aneurysm shapes synthetically deformed into 10,660 models, with 85,280 CFD simulation results and segmentation masks.\n* **Usage Example:** Benchmarking CFD parameter estimation or augmenting data synthetically.\n\n---\n\n**Important:**\n\n1. Only **publicly available, free** datasets are allowed.\n2. List dataset name, URL, and license in your forum post by October 14, 2025.\n3. Use of private or subscription-based data is prohibited.\n\nCombining these datasets for pretraining, augmentation, and multimodal analysis can help boost the accuracy of your aneurysm detection models.\n", "votes": 16 } ], "comments": [], "raw_markdown_by_id": { "3265360": "## 1. Intracranial Aneurysm CTA Images and 3D Models Dataset\n\n* **Description:** High-resolution CTA scans from 99 patients (105 aneurysm cases) plus 44 healthy controls. Includes 3D mesh models of aneurysms and parent vessels, as well as CFD-derived morphological and hemodynamic parameters.\n* **Usage Example:** Pretrain a risk-prediction model by reconstructing CTA into 3D and using CFD features.\n\n## 2. Time-of-Flight MRA of Intracranial Aneurysms with Interval Surveillance Imaging\n\n* **Description:** TOF-MRA scans from 63 patients (24 with longitudinal follow-up). Comes with vessel and aneurysm segmentation masks, clinical annotations, 3D models, and a 3D Slicer project.\n* **Usage Example:** Time-series analysis for aneurysm growth prediction or additional training data for segmentation models.\n\n## 3. Royal Brisbane TOF-MRA Intracranial Aneurysm Database\n\n* **Description:** TOF-MRA scans of 63 patients (85 aneurysm cases) with voxel-level segmentations and expert annotations.\n* **Usage Example:** Fine-grained segmentation training using 3D CNNs or weakly supervised detection.\n\n## 4. Lausanne UIA TOF-MRA Dataset (Di Noto et al., 2022)\n\n* **Description:** 284 subjects (157 aneurysm cases, 127 healthy controls). Includes 246 coarse (spherical) labels and 38 detailed labels—ideal for mixed weak/strong supervision.\n* **Usage Example:** Benchmarking weakly supervised learning frameworks.\n\n## 5. ADAM Dataset for UIA Segmentation (Timmins et al.)\n\n* **Description:** 113 subjects (93 unruptured intracranial aneurysm cases, 20 healthy). Provides voxel-level segmentation masks focused on central cerebral vessels.\n* **Usage Example:** External validation set for transfer-learning and generalization studies.\n\n## 6. IntrA: 3D Intracranial Aneurysm Dataset for Deep Learning\n\n* **Description:** Point-cloud and mesh representations of intracranial aneurysms with classification and part-segmentation benchmarks.\n* **Usage Example:** Training point-cloud networks or mesh-CNNs to learn shape features.\n\n## 7. Aneumo: Large-Scale Multimodal Aneurysm CFD Dataset\n\n* **Description:** 427 real aneurysm shapes synthetically deformed into 10,660 models, with 85,280 CFD simulation results and segmentation masks.\n* **Usage Example:** Benchmarking CFD parameter estimation or augmenting data synthetically.\n\n---\n\n**Important:**\n\n1. Only **publicly available, free** datasets are allowed.\n2. List dataset name, URL, and license in your forum post by October 14, 2025.\n3. Use of private or subscription-based data is prohibited.\n\nCombining these datasets for pretraining, augmentation, and multimodal analysis can help boost the accuracy of your aneurysm detection models.\n" } }