Efficient slice anomaly detection network for 3D brain MRI Volume

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 contrastive loss 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://github.com/Jarvisarmy/SimpleSliceNet.Author summary: In this work, we address the need for more effective methods to identify unusual patterns in brain MRI scans, which are crucial for diagnosing neurological conditions. Traditional techniques often struggle with the high diversity in normal brain appearances and are usually slow and demanding on computing resources. To overcome these challenges, we developed SimpleSliceNet, a streamlined approach that simplifies the analysis without sacrificing accuracy. Our method operates more quickly and requires fewer computational resources than older methods, making it highly practical for clinical settings. By improving the detection of anomalies in brain images, our method holds promise for earlier and more accurate diagnoses of neurological issues, which can directly enhance patient care. Additionally, our approach reduces the technical barriers often faced in medical imaging, potentially broadening the access to high-quality diagnostic techniques across different healthcare environments. Our findings demonstrate that our approach surpasses current leading methods, suggesting a significant step forward in the field of medical imaging.