Skip to main content
. 2017 Sep 4;17(9):2022. doi: 10.3390/s17092022

Figure 2.

Figure 2

Flow chart of the deep meta-architecture approach used in this work. Our system proposes to treat a deep meta-architecture as an open system on which different feature extractors can be adapted to perform on our task. The system is trained and tested end-to-end using images captured in-place. The outputs are the class and localization of the infected area in the image.