Considering the amount of gate counts in implementing all
the functional blocks which could be more than 10 million
gates, and design verification cost with deep sub-micron
process such as 65nm, single chip solution could be
burdensome when quick re-spin is required to support new
features from the market. Assuming that more than half of the
gate count are used for image enhancement processing,
developing a separate chip covering only functional blocks
(left side of Fig. 3) could be more economical and easier to
support new features from the market.