Skip to main content
. 2017 May 22;8:15037. doi: 10.1038/ncomms15037

Figure 3. Image feature decoding and the homology of CNN and the human brain.

Figure 3

(a) Example of a feature unit. True (black) and predicted (light grey) feature values for 50 test images are shown for Unit #562 of CNN Layer 8 (CNN8) predicted from the whole VC. (b) Image feature decoding accuracy. Mean decoding accuracies are shown for each combination of the feature type/layer and ROI (error bars, 95% confidence interval (CI) across five subjects).