If a "Three-Value" Cell can see two other "Bi-Value" Cells that contain different combinations of the candidates of the first Cell, then the candidate common to the two "Bi-Value" Cells can not be the solution in any Cell that "sees" these three Cells.

Indeed, the "Three-Value" Cell and the two "Bi-Value" Cells form a group of three Cells with exactly three possible candidates, so their solution each contain one of the three candidates, in particular the "common" candidate. Hence the "common" candidate can not be the solution in any Cell that sees the three Cells.

In the above example if candidate 1 were the solution in J3, then there would be only two possible candidates left (3 and 7) for three cells, which is not possible.

You can practice this strategy by installing the SudokuCoach application on your Android™ device.