Skip to main content
. 2018 Feb 5;8:2369. doi: 10.1038/s41598-018-20521-y

Figure 5.

Figure 5

Single-Pixel Camera Performance Results. Comparing the full Hadamard basis (M = 16834 patterns) (a), with the reduced optimised Hadamard basis (M = 666 patterns, 4% of total) (b), the reduced optimised Hadamard basis smoothed using a Gaussian filter with 1 pixel standard deviation (M = 666 patterns) (c) and the deep learned basis (M = 666 patterns) (d). Note the improved sloping lines in DL compared to FHT.