To find solutions of total variation (TV) regularized optimization problems, we want to make use of the Frank Wolfe (FW) algorithm. Therefore, we analyze what the steps of FW look like in this application and develop the structure of solutions.