Filling the gap: the missing eclipses of $γ$ Persei from 2005 and from 2006

$γ$ Persei is a long-period ($P \approx 14.6$ yr) eclipsing binary system. Its period makes it a difficult target to fully understand: so far, only two primary eclipses are known in the literature, from 1990 and from 2019, whereas the 2005 one was missed due to its closeness to the Sun at the time. We aimed to fill in this gap by processing the quasi-continuous photometry collected by the Solar Mass Ejection Imager (SMEI) between 2003 and 2011, which was ideally positioned to observe such a bright targets. In order to do that, we first determined a color-dependent conversion formula from the SMEI measurements into {\it Gaia} \textit{G} magnitudes. We applied various corrections to the photometry and provide the longest continuous light curve of $γ$ Persei. We successfully detected the 2005 primary eclipse of the system, with the yearly observations ending during the egress of the companion. We predicted the position of a possible secondary eclipse by forward modeling the binary system with PHOEBE, and successfully recovered the secondary eclipse in the 2006 SMEI observations. The existence of the secondary eclipse puts strong constraints on the orbital configuration, which will be an important constraint for future studies of the system.