Characterization of diabetic peripheral neuropathy in infrared video sequences using independent component analysis

Diabetic peripheral neuropathy (DPN) is a complication of diabetes that leads to foot ulceration and amputations. Detection of DPN at an early stage is important for clinical management of the disease. This paper presents characterization of DPN subjects using an infrared imaging device and independent component analysis (ICA). Infrared video sequences are captured after cold provocation of the plantar foot. After the video frames are preprocessed and registered, ICA is performed. The objective of using ICA is to separate the temporal and spatial components that may represent thermo regulatory differences in response to a cold provocation between normal controls and patients with DPN. The independent components of control and DPN subjects are compared and tested for statistically significant differences to the temporal responses to cold stimulus. Our results show 3 independent components that differentiate controls and DPN patients with substantial significance.