Limitations and Future Trends
Xiaofeng Yang · 2023
Recent years have witnessed the trend of deep learning being increasingly used in the application of medical imaging. The latest networks and techniques have been borrowed from the field of computer vision and adapted to specific clinical tasks in radiology and radiation oncology. With further development in both artificial intelligence and computing hardware, more learning-based methods are expected to facilitate the clinical workflow with novel applications. Although the reviewed literature shows the success of deep learning-based image synthesis in various applications, there remain some open questions to be answered in future studies. Unlike conventional methods, learning-based methods require large training datasets. The size of training sets has been shown to affect the performance of machine learning in many challenging computer vision problems as well as medical imaging tasks. Generally, a larger training set size with greater data variation can reduce overfitting of the model and enable better performance.