Efficient Slice Anomaly Detection Network for 3D Brain MRI Volume

TL;DR

This paper introduces SimpleSliceNet, a memory-efficient and accurate anomaly detection framework for 3D brain MRI data, leveraging pre-trained 2D slice features and advanced probabilistic modeling.

Contribution

It presents a novel slice-based network that combines pre-trained 2D feature extraction with normalizing flows and a semi-push-pull mechanism for improved anomaly detection in 3D MRI volumes.

Findings

Outperforms state-of-the-art models in accuracy, memory, and speed.

Reduces computational cost by using 2D slice features.

Demonstrates high adaptability to brain MRI anomaly detection.

Abstract

Current anomaly detection methods excel with benchmark industrial data but struggle with natural images and medical data due to varying definitions of 'normal' and 'abnormal.' This makes accurate identification of deviations in these fields particularly challenging. Especially for 3D brain MRI data, all the state-of-the-art models are reconstruction-based with 3D convolutional neural networks which are memory-intensive, time-consuming and producing noisy outputs that require further post-processing. We propose a framework called Simple Slice-based Network (SimpleSliceNet), which utilizes a model pre-trained on ImageNet and fine-tuned on a separate MRI dataset as a 2D slice feature extractor to reduce computational cost. We aggregate the extracted features to perform anomaly detection tasks on 3D brain MRI volumes. Our model integrates a conditional normalizing flow to calculate log likelihood of features and employs the Semi-Push-Pull Mechanism to enhance anomaly detection accuracy. The results indicate improved performance, showcasing our model's remarkable adaptability and effectiveness when addressing the challenges exists in brain MRI data. In addition, for the large-scale 3D brain volumes, our model SimpleSliceNet outperforms the state-of-the-art 2D and 3D models in terms of accuracy, memory usage and time consumption. Code is available at: https://anonymous.4open.science/r/SimpleSliceNet-8EA3.