The accurate detection and quantification of human diabetic somatic polyneuropathy (DSPN) are important to define at risk patients, anticipate deterioration, and assess new therapies. Current methods lack sensitivity, require expert assessment and have major shortcomings when employed to define therapeutic efficacy. In recent years, in vivo corneal confocal microscopy (IVCCM) has shown potential as a surrogate endpoint for DSPN.This study aims to investigate fundamental aspects of IVCCM such as repeatability and optimal scanning methodology and establish changes in corneal nerve morphology in relation to the severity of DSPN and regeneration in response to normalisation of hyperglycaemia. Furthermore, it aims to provide a novel fully automated image analysis algorithm for the quantification of corneal nerve morphology and establish the diagnostic ability of CCM.IVCCM shows high repeatability which is enhanced with more experienced observers. Central corneal innervation is comparable to adjacent peripheral innervation in mild diabetic neuropathy but the central cornea may be more sensitive to change. Corneal nerve loss is symmetrical and progressive with increasing neuropathic severity and corneal nerves show significant regenerative capacity following rapid normalisation of glycaemic control after simultaneous pancreas and kidney transplantation. The novel image analysis algorithm strongly correlates with human expert annotation and therefore represents a rapid, objective and repeatable means of assessing corneal nerve morphology. Automated image quantification may replace human manual assessment with high diagnostic validity for DSPN.