Two data modalities, ultrasound images and numeric clinical records, are categorized into Thrombus and Non-Thrombus classes for lower limb diagnosis. The dataset is sourced from the ThrombUS+ clinical site to facilitate automated deep vein thrombosis detection.
Use Cases
- Train a multimodal classifier using ultrasound images and numeric features to identify lower limb thrombus.
- Analyze the numeric clinical data to determine correlations with the Thrombus/Non-Thrombus labels.
- Develop image-based diagnostic algorithms for deep vein thrombosis using the provided ultrasound scans.
Strengths
- Includes ultrasound images and numeric clinical data for multimodal analysis.
- Provides binary classification labels for Thrombus and Non-Thrombus cases.
- Sourced from the ThrombUS+ clinical validation study.