TriALS is a multi-centre benchmark dataset for liver lesion segmentation in non-contrast CT scans. It contains images from the same patient across four CT phases, including non-contrast, arterial, portal venous, and delayed phases, with lesion annotations. The dataset was created by marwankefah and last updated on 2026-05-19.
Use Cases
- Developing segmentation models for liver lesions based on multi-phase CT images.
- Benchmarking algorithm performance on occult lesions that become visible only after contrast administration.
- Training AI systems to identify lesions in non-contrast CT scans, a core diagnostic challenge mentioned in the description.
Strengths
- Contains images from the same patient across four distinct CT phases, providing aligned multi-phase data.
- Includes lesion annotations, enabling supervised learning tasks.
- Described as the first multi-centre benchmark for liver lesion segmentation in non-contrast CT.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- huggingface
- Freshness
- Last updated 2026-05-19 13:34:35; freshness should be verified.