A comparison of ImageJ and machine learning based image analysis methods to measure cassava bacterial blight severity

This dataset contains raw images and results from image analysis of cassava leaves infiltrated with Xanthomonas, the causal agent of cassava bacterial blight, a Xanthomonas mutant with decreased pathogenecity, and mock treatments. Leaves were imaged using a Raspberry Pi camera system at 0, 4, 6, and 9 days post inoculation. Disease symptoms on the leaf known as "watersoaked lesions" were segmented and analyzed using two different image analysis methods. First, the open-source software ImageJ and second, a machine learning tool developed in our lab. Both programs were compared for their ability to segment the lesions and distinguish between lesion types.