wherein a current block in the image is divided into two regions including a first region and a second region,

the method comprising:

generating first prediction samples in the first region in the current block based on a first reference sample group for the first region, reference samples in the first reference sample group being located outside the current block; and

generating second prediction samples in the second region in the current block based on a second reference sample group for the second region, the second reference sample group comprising at least one reference sample located outside the current block and at least one first prediction sample in the first region,

wherein a first prediction sample in the first region is generated based on an averaged value of a plurality of reference samples in the first reference sample group,

wherein a second prediction sample in the second region is generated by utilizing either a vertical reference sample or a horizontal reference sample in the second reference sample group, the vertical reference sample being lying on the same vertical line as the second prediction sample and the horizontal reference sample being lying on the same horizontal line as the second prediction sample,

wherein third prediction samples in the second region are further generated based on the second prediction samples, that are generated after the first prediction samples, in the second region,

wherein a third prediction sample in the second region is derived by applying a pre-defined filter to a plurality of second prediction samples in the second region, and

wherein the plurality of second prediction samples for deriving the third prediction sample are on the same vertical or horizontal line.