Fig. 3 | Scientific Reports

Fig. 3

From: Convolutional neural networks for accurate real-time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high-resolution in vivo confocal microscopy

Fig. 3The alternative text for this image may have been generated using AI.

Flow diagram for workflow of CNN quality and diagnostic analysis of in vivo captured micrographs. This flow diagram shows the processing workflow of raw in vivo images captured by the InVivage confocal endomicroscope. The raw images were first used to develop the quality filtering CNN (QMR). Following this the QMR was used to filter the entire dataset (n = 9168). The diagnostic quality images after filtering (n = 1983) were divided based on contrast agent used and assigned to the acriflavine diagnostic CNN (APMAC) (n = 1343) and fluorescein diagnostic CNN (FMPAC) (n = 640). These diagnostic CNNs were developed using these images to classify them into ‘no dysplasia’, ‘lichenoid’, ‘low-risk’, and ‘high-risk’.

Back to article page