In recent years, considerable work has been conducted on the development of synthetic medical images, but there are no satisfactory methods for evaluating their medical suitability. Existing methods mainly evaluate the quality of noise in the images, and the similarity of the images to the real images used to generate them. For this purpose, they use feature maps of images extracted in different ways or distribution of images set. Then, the proximity of synthetic images to the real set is evaluated using different distance metrics. However, it is not possible to determine whether only one synthetic image was generated repeatedly, or whether the synthetic set exactly repeats the training set. In addition, most evolution metrics take a lot of time to calculate. Taking these issues into account, we have proposed a method that can quantitatively and qualitatively evaluate synthetic images. This method is a combination of two methods, namely, FMD and CNN-based evaluation methods. The estimation methods were compared with the FID method, and it was found that the FMD method has a great advantage in terms of speed, while the CNN method has the ability to estimate more accurately. To evaluate the reliability of the methods, a dataset of different real images was checked.

The generation of synthetic medical data has become a focal point for researchers, driven by the increasing demand for privacy-preserving solutions. While existing generative methods heavily rely on real datasets for training, access to such data is often restricted. In contrast, statistical information about these datasets is more readily available, yet current methods struggle to generate tabular data solely from statistical inputs. This study addresses the gaps by introducing a novel approach that converts statistical data into tabular datasets using a modified Generative Adversarial Network (GAN) architecture. A custom loss function was incorporated into the training process to enhance the quality of the generated data. The proposed method is evaluated using fidelity and utility metrics, achieving “Good” similarity and “Excellent” utility scores. While the generated data may not fully replace real databases, it demonstrates satisfactory performance for training machine-learning algorithms. This work provides a promising solution for synthetic data generation when real datasets are inaccessible, with potential applications in medical data privacy and beyond.