Towards an Automated Analysis of Forearm Thermal Images During Handgrip Exercise

Infrared thermography is an imaging modality that is able to map the surface skin temperature and has been widely researched and applied in biomedical applications. It can be used as a monitoring method for upper limb condition support diagnosis and treatment assessment. The aim of this research is the development of an image processing method for rapid analysis of forearm thermographic images. The images were taken from 13 participants who undergone in a handgrip force exercise using a FLIR A325sc thermal camera. The obtained images were analyzed with FLIR ThermaCAM Researcher Pro 2.10 software package, and by a customer developed program in Matlab 2016b. Three regions of interest (ROI) were defined for both analysis. At the Matlab developed application, the forearm was segmented from the background and the ROIs placed automatically and average temperature and standard deviation were extracted. The results from both analysis were compared and showed good correlations for the examined 1365 images. The Matlab developed analysis is much faster than the time-consuming analysis performed using the standard camera manufacturer provided tool, however the first has to be used to export the temperature matrix from the proprietary image format in order to be used by the Matlab developed program. The development of a reliable tool for automated thermographic analysis of the forearms ROIs was successful implemented. For further work, it is proposed to read directly from the raw thermal images proprietary formats to hasten the analysis.