Medical image classification based on artificial intelligence approaches

摘要: A new intelligent system to tackle the main challenges of confocal corneal imaging is developed.This underpins expertise ophthalmologists.It provides clinically useful factors, saves a amount clinician time in process.It able model stromal keratocyte cells for better evaluation and fast analysis.Early approval by clinicians. Corneal images can be acquired using microscopes which provide detailed views different layers inside human cornea. Some problems diseases occur one or more layers: epithelium, stroma endothelium. Consequently, automatically extracting clinical information associated with diseases, identifying abnormality evaluating normal cornea, it important recognise these reliably. Artificial intelligence (AI) approaches improved accuracy over conventional processing techniques save manual analysis required experts. neural networks (ANNs), adaptive neuro fuzzy inference systems (ANFIS) committee machine (CM) have been investigated tested improve recognition identify layers. The performance CM, formed from ANN ANFIS, achieves an 100% some classes processed data sets. Three sets seven abnormal proposed system. Statistical performed track any change images. This pre-process (quality enhancement, noise removal), classify images, abnormalities analysed visualise as well each individual cell 3D volume further analysis.