he invention relates to a method for encoding at least one image broken up into blocks, implementing, for at least one current block, a step of transforming (11) said current block from the spatial domain into a transformed domain, outputting a transformed block consisting of a set of coefficients representative of said current block, referred to as original coefficients. According to the invention, such a method implements a step of selecting (12) a subset of original coefficients from said set and, for at least one original coefficient of said subset, the following steps: estimating (13) said original coefficient, outputting an estimated coefficient in said transformed domain; determining (14) a residual of a coefficient by comparing said original coefficient and said estimated coefficient; and encoding (15) said residual of a coefficient.