In-Domain Transfer Learning Strategy for Tumor Detection on Brain MRI.

Transfer learning has gained importance in areas where there is a labeled data shortage. However, it is still controversial as to what extent natural image datasets as pre-training sources contribute scientifically to success in different fields, such as medical imaging. In this study, the effect of transfer learning for medical object detection was quantitatively compared using natural and medical image datasets. Within the scope of this study, transfer learning strategies based on five different weight initialization methods were discussed. A natural image dataset MS COCO and brain tumor dataset BraTS 2020 were used as the transfer learning source, and Gazi Brains 2020 was used for the target. Mask R-CNN was adopted as a deep learning architecture for its capability to effectively handle both object detection and segmentation tasks. The experimental results show that transfer learning from the medical image dataset was found to be 10% more successful and showed 24% better convergence performance than the MS COCO pre-trained model, although it contains fewer data. While the effect of data augmentation on the natural image pre-trained model was 5%, the same domain pre-trained model was measured as 2%. According to the most widely used object detection metric, transfer learning strategies using MS COCO weights and random weights showed the same object detection performance as data augmentation. The performance of the most effective strategies identified in the Mask R-CNN model was also tested with YOLOv8. Results showed that even if the amount of data is less than the natural dataset, in-domain transfer learning is more efficient than cross-domain transfer learning. Moreover, this study demonstrates the first use of the Gazi Brains 2020 dataset, which was generated to address the lack of labeled and qualified brain MRI data in the medical field for in-domain transfer learning. Thus, knowledge transfer was carried out from the deep neural network, which was trained with brain tumor data and tested on a different brain tumor dataset.

A comprehensive survey of deep learning research on medical image analysis with focus on transfer learning.

This survey aims to identify commonly used methods, datasets, future trends, knowledge gaps, constraints, and limitations in the field to provide an overview of current solutions used in medical image analysis in parallel with the rapid developments in transfer learning (TL). Unlike previous studies, this survey grouped the last five years of current studies for the period between January 2017 and February 2021 according to different anatomical regions and detailed the modality, medical task, TL method, source data, target data, and public or private datasets used in medical imaging. Also, it provides readers with detailed information on technical challenges, opportunities, and future research trends. In this way, an overview of recent developments is provided to help researchers to select the most effective and efficient methods and access widely used and publicly available medical datasets, research gaps, and limitations of the available literature.

Structural profile matrices for predicting structural properties of proteins.

Predicting structural properties of proteins plays a key role in predicting the 3D structure of proteins. In this study, new structural profile matrices (SPM) are developed for protein secondary structure, solvent accessibility and torsion angle class predictions, which could be used as input to 3D prediction algorithms. The structural templates employed in computing SPMs are detected by eight alignment methods in LOMETS server, gap affine alignment method, ScanProsite, PfamScan, and HHblits. The contribution of each template is weighted by its similarity to target, which is assessed by several sequence alignment scores. For comparison, the SPMs are also computed using Homolpro, which uses BLAST for target template alignments and does not assign weights to templates. Incorporating the SPMs into DSPRED classifier, the prediction accuracy improves significantly as demonstrated by cross-validation experiments on two difficult benchmarks. The most accurate predictions are obtained using the SPMs derived by threading methods in LOMETS server. On the other hand, the computational cost of computing these SPMs was the highest.

Developing structural profile matrices for protein secondary structure and solvent accessibility prediction.

MOTIVATION: Predicting secondary structure and solvent accessibility of proteins are among the essential steps that preclude more elaborate 3D structure prediction tasks. Incorporating class label information contained in templates with known structures has the potential to improve the accuracy of prediction methods. Building a structural profile matrix is one such technique that provides a distribution for class labels at each amino acid position of the target. RESULTS: In this paper, a new structural profiling technique is proposed that is based on deriving PFAM families and is combined with an existing approach. Cross-validation experiments on two benchmark datasets and at various similarity intervals demonstrate that the proposed profiling strategy performs significantly better than Homolpro, a state-of-the-art method for incorporating template information, as assessed by statistical hypothesis tests. AVAILABILITY AND IMPLEMENTATION: The DSPRED method can be accessed by visiting the PSP server at http://psp.agu.edu.tr. Source code and binaries are freely available at https://github.com/yusufzaferaydin/dspred. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.