To respond to an image large in the moving amount of input image data by achieving high encoding efficiency of a moving image.
In this device 100 for encoding a moving image, an inter-screen prediction part 108 generates an inter-screen prediction image signal InterPV by a motion retrieval/motion compensation process, and an in-screen prediction part 109 generates an in-screen prediction image signal IntraPV. An in-screen/inter-screen prediction image selection part 110 selects an image high in encoding efficiency from the InterPV and IntraPV to output a prediction image Used PV, and forms an inter-screen prediction error signal InterAD and an in-screen prediction error signal IntraAD. An image feature amount estimation part 111 generates an inter-screen prediction error average value InterSAD and an in-screen prediction error average value IntraSAD, and calculates an image feature amount V of the difference thereof. The image feature amount V is supplied to an image input part 102, and the insertion number of bidirectional prediction encoded pictures is controlled.
YOKOYAMA TORU
NAKAMURA KATSUYUKI
IMAOKA REN
MOCHIZUKI SEIJI