Balanced frames: a useful tool in signal processing with good properties (Q2210810)

The authors consider balanced unit norm tight frames in a \(d\)-dimensional Hilbert space \({\mathbb H}_d.\) A frame \((f_k)_{k=1}^K\) for \({\mathbb H}_d\) is said to be balanced if \(\sum_{k=1}^K f_k = 0.\) After a preliminary section on basic results of frame theory, section 3 describe several advantages of balanced frames for applications. For instance, the reconstructions using balanced frames are robust against systematic errors in the frame coefficients. Section 4 present some versions of Naimark's theorem for balanced frames (theorems 4.8 and 4.10). Theorem 5.5 gives a necessary and sufficient condition for the existence of a closest balanced frame to a given frame, both in the \(\ell^1\) norm and in the \(\ell^2\) norm. Section 6 discusses a new concept of complement for balanced frames and the final sections 7 and 8 give explicit examples of balanced unit norm tight frames and study some of their properties.