This study aims to present a new idea to expand the perception area of each cell in cellular automaton (CA) using a recursive algorithm known as Recursive Estimation of Neighbors. An intelligent cellular process defined by the algorithm makes it possible to introduce an extra radius of the perception area, in addition to the radius of the CA neighborhood. A basic CA rule is extrapolated into rules with larger radii, which form a sequence indexed by the extra radius containing the basic CA as the first term of the sequence. The patterns formed in some typical sequences of extrapolated ECA and Life-like CA rules are presented. Contrasting pattern activities contained in homogeneous and heterogeneous CAs are discussed by applying mean field analysis. Some symmetrical arrangements of composites of cells with different radii are used in order to discuss the heterogeneous CA. The new perspective presented here offers several possible applications for CA.