LiMT: A Multi-task Liver Image Benchmark Dataset

2025-11-24

AI Summary

Existing liver imaging datasets are predominantly designed for single-task learning, limiting their utility for multi-task collaborative computer-aided diagnosis (CAD). Method: We introduce LiverMT, the first publicly available multi-task CT dataset specifically curated for liver CAD. It comprises 150 arterial-phase contrast-enhanced CT scans covering normal livers and four common hepatopathologies. All scans are meticulously annotated by clinical experts for three complementary tasks: liver and tumor segmentation, multi-label lesion classification, and lesion detection—under a unified annotation protocol. Contribution/Results: LiverMT explicitly models inter-task relationships and mitigates data heterogeneity across tasks. We provide benchmark multi-task baseline models and a systematic literature review, establishing the first standardized multi-task liver imaging benchmark. This resource advances research in cross-task joint modeling and facilitates holistic, clinically grounded liver CAD development.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multi-instance/Multi-view Learning
• Data Mining & Knowledge Management: Mining of Visual, Multimedia & Multimodal Data

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data

Abstract

Computer-aided diagnosis (CAD) technology can assist clinicians in evaluating liver lesions and intervening with treatment in time. Although CAD technology has advanced in recent years, the application scope of existing datasets remains relatively limited, typically supporting only single tasks, which has somewhat constrained the development of CAD technology. To address the above limitation, in this paper, we construct a multi-task liver dataset (LiMT) used for liver and tumor segmentation, multi-label lesion classification, and lesion detection based on arterial phase-enhanced computed tomography (CT), potentially providing an exploratory solution that is able to explore the correlation between tasks and does not need to worry about the heterogeneity between task-specific datasets during training. The dataset includes CT volumes from 150 different cases, comprising four types of liver diseases as well as normal cases. Each volume has been carefully annotated and calibrated by experienced clinicians. This public multi-task dataset may become a valuable resource for the medical imaging research community in the future. In addition, this paper not only provides relevant baseline experimental results but also reviews existing datasets and methods related to liver-related tasks. Our dataset is available at https://drive.google.com/drive/folders/1l9HRK13uaOQTNShf5pwgSz3OTanWjkag?usp=sharing.

Problem

• Existing liver datasets are limited to single tasks, constraining CAD development
• LiMT provides multi-task data for segmentation, classification and lesion detection
• The dataset addresses heterogeneity issues between task-specific datasets during training

Innovation

• Multi-task liver dataset for segmentation classification detection
• Arterial phase CT volumes with clinician annotations
• Explores task correlations without dataset heterogeneity issues