To correctly decide a compression parameter by which a target compression rate is attained by only a few test compression times by calculating the proper compression parameter to obtain the target compression rate and compressing image data through the use of the calculated compression parameter.

A compression processing part 18 executes communication with a control part 21 and obtains compression image quality setting by an operation button group 24. The part 18 decides the target compression rate in accordance with the image quality setting. Then approximation is obtained by a relational expression: Y=g/(X+c)+h (g, c and h are an unfixed parameter) so that 'relation between the compression parameter X and a compression code amount Y' which indicates an inverted S shape curve is excellently approximated within a wide range. The compression parameter X which seems to be proper in attaining the target compression rate is correctly obtained by the inverse calculation of the relational expression. Then the test compression times are efficiently reduced.