We study the Pareto frontier for two competing norms · X and · Y on a vector space. For a given vector c, the Pareto frontier describes the possible values of ( a X , b Y ) for a decomposition c = a + b. The singular value decomposition of a matrix is closely related to the Pareto frontier for the spectral and nuclear norm. We will develop a general theory that extends the notion of singular values of a matrix to arbitrary finite dimensional euclidean vector spaces equipped with dual norms. This also generalizes the diagonal singular value decompositions for tensors introduced by the author in previous work. We can apply the results to denoising, where c is a noisy signal, a is a sparse signal and b is noise. Applications include 1D total variation denoising, 2D total variation Rudin-Osher-Fatemi image denoising, LASSO, basis pursuit denoising and tensor decompositions. arXiv:1705.10881v1 [math.NA] 30 May 2017 4 12. Total Variation Denoising in Imaging 48 12.1. The 2D total variation norm 48 12.2. Spareseness and total variation 49 12.3. The total variation norm is not tight 50 13. Tensor Decompositions 50 13.1. CP decompositions 50 13.2. The CoDe model and the nuclear norm 51 13.3. Examples of unitangent tensors 51 13.4. The diagonal SVD and the slope decomposition 52 13.5. Group algebra tensors 54 13.6. symmetric tensors in R 2×2×2 54 References 57 2. Main Results 2.1. The Pareto frontier. Let us consider a finite dimensional R-vector space V equipped with two norms, · X and · Y .Suppose that c ∈ V . We are looking for decompositions c = a + b that are optimal in the sense that we cannot reduce a X without increasing b Y and we cannot reduct b Y without increasing a X . We recall the definition from the introduction:there exists a decomposition c = a + b with a X = x, b Y = y such that for every decomposition c = a + b we have a X > x, b Y > y or ( a X , b Y ) = (x, y). If (x, y) is a Pareto efficient pair then we call c = a + b an XY -decomposition. By symmetry, c = a + b is an XY -decomposition if and only if c = b + a is a Y Xdecomposition. The Pareto frontier consists of all Pareto efficient pairs (see [6]). The Pareto frontier is the graph of a strictly decreasing, continuous convex function f c Y X : [0, c X ] → [0, c Y ] (see [6] and Lemmas 3.2 and 3.3). If we change the role of X and Y we get the graph of f c XY , so f c XY and f c Y X are inverse functions of each other. Example 2.2. Consider the vector space V = R n . Sparseness of vectors in R n can be measured by the number of nonzero entries. For c ∈ V we defineNote that · 0 is not a norm on V because it does not satisfy λc 0 = |λ| c 0 for λ ∈ R.Convex relaxation of · 0 gives us the 1 norm · 1 . This means that the unit ball for the norm · 1 is the convex hull of all vectors c with c 0 = c 2 = 1. Let us take · X = · 1 and · Y = · ∞ and describe the Pareto frontier. Suppose that c = (c 1 · · · c n ) t ∈ R n and 0 ≤ y ≤ c ∞ . If b ∞ = y then we have