Image encoding and decoding method with merge flag and motion vectors

An image decoding method of dividing an input image signal into pixel blocks, and performing inter prediction on divided pixel blocks, comprising:

• deriving a plurality of motion information candidates from neighboring blocks of a target block in a first frame and a collocated block in a second frame different from the first frame, wherein the collocated block is selected according to a center position of the target block or a corner position of the target block, and information specifying whether the collocated block is selected according to the center position of the target block or the corner position of the target block is derived based on at least one of a position and a size of the target block;
• decoding an index for specifying a merging block from the neighboring blocks and the collocated block, wherein the index is not decoded when the number of motion information candidates is one;
• selecting predicted motion information from the plurality of motion information candidates in accordance with the index;
• and
• predicting motion information of the target block by using the predicted motion information.