Semi-supervised VAE-GAN for Out-of-Sample Detection Applied to MRI Quality Control

Recent Advances in Variational Autoencoders With Representation Learning for Biomedical Informatics: A Survey

“…VAEs have also succeeded in biological image analyses, and many studies show superior performance. The main research area based on the VAE use in medical imaging datasets includes: 1) Medical image data augmentation for downstream tasks include image classification [68,71,79,80], image segmentation [72][73][74][75][76]87], image restoration [85,86], and image reconstruction [72,[81][82][83].…”

“…Gomez et al [82] proposed an image reconstruction architecture based on β -VAE, which can combine numerous overlapping image patches into a fusion reconstruction of the real fetal ultrasound images. Mostapha et al [83] proposed VAEGANbased framework to the automatic quality control of structural MR images. Tudosiu et al [84] proposed a model based on VQ-VAE, which can effectively encode full-resolution 3D brain volume, compressing data to 0.825% of the original size, while maintaining image fidelity.…”

Applications of Generative Adversarial Networks in Anomaly Detection: A Systematic Literature Review

“…Anomaly detection in medicine deals with analyzing patients' health conditions using medical records and images [153]. Specific applications include retinal optical coherence tomography (OCT) anomaly detection [17,26,29,104], seizure detection [18], cardiovascular disease detection [30], lung nodule detection [42], abnormal chest X-ray identification [59,138,97,135], polyp detection [123,80], metastatic bone tumor detection [78], lesion detection [137,101], laparoscopy anomaly detection [85], breast cancer detection [143,132], MRI quality control [98], diabetic retinopathy detection [133], brain tumor detection [134] and hemorrhage detection [105]. One of the challenges in this domain is the difficulty of obtaining expert labels for medical data, such as clinical images, since annotation is an exhaustive and time-consuming task.…”

“…SSIM, adopted by seven papers [26,55,56,63,65,98,99], quantifies the relative perceptual similarity between two images. This metric ranges from -1 to 1, with 1 indicating a perfect pixel match between the original and generated samples, -1 corresponding to inverted images, and 0 marking no similarity [190].…”

“…Structural similarity indices metrics (SSIM): [116,55,56,63,65,98,99] Peak signal to noise ratio (PSNR): [21,22,24,41,44,137,85] Visual inspection: [17,40,60,69,139,135] Fréchet inception distance (FID): [124,82] Signal to noise ratio (SNR): [53] L2-norm distance: [109] Fully convolutional network (FCN)-score: [65] Earth mover's distance: [21] Cosine similarity: [109] [116,55,56,63,65,98,99] Peak signal to noise ratio (PSNR): [21,22,24,41,44,137,85] Visual inspection: [17,40,60,69,139,135] Fréchet inception distance (FID): [124,…”