DeepDicomSort: An Automatic Sorting Algorithm for Brain Magnetic Resonance Imaging Data

Abstract

With the increasing size of datasets used in medical imaging research, the need for automated data curation is arising. One important data curation task is the structured organization of a dataset for preserving integrity and ensuring reusability. Therefore, we investigated whether this data organization step can be automated. To this end, we designed a convolutional neural network (CNN) that automatically recognizes eight different brain magnetic resonance imaging (MRI) scan types based on visual appearance. Thus, our method is unaffected by inconsistent or missing scan metadata. It can recognize precontrast T1-weighted (T1w), post-contrast T1-weighted (T1wC), T2-weighted (T2w), proton density-weighted (PDw) and derived maps (e.g. apparent diffusion coefficient and cerebral blood flow). In a first experiment, we used scans of subjects with brain tumors: 11065 scans of 719 subjects for training, and 2369 scans of 192 subjects for testing. The CNN achieved an overall accuracy of 98.7%. In a second experiment, we trained the CNN on all 13434 scans from the first experiment and tested it on 7227 scans of 1318 Alzheimer's subjects. Here, the CNN achieved an overall accuracy of 98.5%. In conclusion, our method can accurately predict scan type, and can quickly and automatically sort a brain MRI dataset virtually without the need for manual verification. In this way, our method can assist with properly organizing a dataset, which maximizes the shareability and integrity of the data.

The Essential Role of Open Data and Software for the Future of Ultrasound-Based Neuronavigation

“…One example is the work of Sagheer et al ( 44 ) who used the BITE database to validate their US image denoising algorithm. Other examples include a US probe calibration method ( 45 ) where the authors used the BITE database in the validation process, simulation of 2D US from 3D MR ( 46 ) and sorting of DICOM images ( 47 ). Open image databases such as BITE and RESECT can also be used for more clinically oriented research.…”

MR-Class: A Python Tool for Brain Mr Image Classification Utilizing One-Vs-All DCNNS to Deal With the Open-Set Recognition Problem

“…In the context of medical image retrieval and classification using DCNNs, three different DCNNs are present, i.e., ResNet-18 [14], Φ-Net [15], and DeepDicomSort [16]. Hence, a comparison study was performed where the architecture that showed the highest classification accuracy was adopted in the one-vs-all training approach.…”

“…After the success of the deep convolutional neural network (DCNN), AlexNet [8] in the ImageNet [9] classification challenge, an increase of interest in DCNN has been seen when dealing with image classification tasks [10][11][12]. In the context of medical image retrieval and classification using DCNNs, four different studies have been identified for the classification of body organs and MR images (Accuracy >90%) [13][14][15][16]. A summary of these models can be seen in Supplementary-Table S1.…”

Review: Deep Learning in Electron Microscopy

“…However, large datasets like DeepMind Kinetics 601 , ImageNet 564 , and YouTube 8M 602 may take a team months to prepare. As a result, it may not be practical to divert sufficient staff and resources to curate a high-quality dataset, even if curation is partially automated [602][603][604][605][606][607][608][609] . To curate data, human capital can be temporarily and cheaply increased by using microjob services 610 .…”