Figure 1
The computational pipeline for employing spatial independent component analysis (sICA) on thermal imaging data (A) Motion correction framework showing co-registered images with and without application of optical flow. For each subject, the whole thermal imaging data are aggregated into a single matrix, in which each row represents the thermal imaging data in one time point and each column stands for the time series of one single pixel (B) A mask to exclude background, i.e. neck and clothes, and retain only the face was applied to each frame. The data from these images were used as the source matrix (C) sICA—illustration of the mixing matrix, each column of which represents the time course of the corresponding source signal. An exemplar time series of nose components is shown. Each row of the source signal matrix represents one independent spatial map. Thermal image of subject L06 is used for illustrative purposes.
