Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Large-Scale 3D Medical Image Pre-training with Geometric Context Priors

About

The scarcity of annotations poses a significant challenge in medical image analysis. Large-scale pre-training has emerged as a promising label-efficient solution, owing to the utilization of large-scale data, large models, and advanced pre-training techniques. However, its development in medical images remains underexplored. The primary challenge lies in harnessing large-scale unlabeled data and learning high-level semantics without annotations. We observe that 3D medical images exhibit consistent geometric context, i.e., consistent geometric relations between different organs, which leads to a promising way for learning consistent representations. Motivated by this, we introduce a simple-yet-effective Volume Contrast (VoCo) framework to leverage geometric context priors for self-supervision. Given an input volume, we extract base crops from different regions to construct positive and negative pairs for contrastive learning. Then we predict the contextual position of a random crop by contrasting its similarity to the base crops. In this way, VoCo encodes the inherent geometric context into model representations, facilitating high-level semantic learning without annotations. Specifically, we (1) introduce the largest medical pre-training dataset PreCT-160K; (2) investigate scaling laws and propose guidelines for tailoring different model sizes to various medical tasks; (3) build a benchmark encompassing 48 medical tasks. Extensive experiments highlight the superiority of VoCo. Codes at https://github.com/Luffy03/Large-Scale-Medical.

Linshan Wu, Jiaxin Zhuang, Hao Chen• 2024

Related benchmarks

TaskDatasetResultRank
Pan-cancer SegmentationInternal datasets
Lung Tumor DSC53.4
14
Pan-cancer ScreeningFLARE 2023
DSC50.6
10
Pan-cancer SegmentationIRCADb liver tumors (External)
DSC0.565
10
Pan-cancer SegmentationRider lung tumors (External)
DSC (%)31.3
10
Pan-cancer SegmentationExternal Datasets Rider, Corona, IRCADb Average
Average DSC (%)47.9
10
Abdominal Organ SegmentationFLARE 23
Duration114
10
Pan-cancer SegmentationCorona COVID-19 (External)
DSC55.8
10
Pan-cancer detectionInternal 16 datasets (val)
Lung Tumor91.1
10
Pan-cancer SegmentationHealthy Datasets CHAOS, TCIA, Atlas
CHAOS Score40
10
Pan-cancer ScreeningMICCAI FLARE25 leaderboard (val)
DSC47.5
4
Showing 10 of 10 rows

Other info

Follow for update