What is odd about binary Parseval frames? (Q1684509)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | What is odd about binary Parseval frames? |
scientific article |
Statements
What is odd about binary Parseval frames? (English)
0 references
11 December 2017
0 references
The present paper is concerned with binary Parseval frames. These binary frames provide explicit expansions of binary vectors using a bilinear form that resembles the dot product in Euclidean spaces. In contrast to the inner product on real or complex Hilbert spaces, there are many nonzero vectors whose dot product with themselves vanishes. Such vectors have special significance in the paper results. Due to the number of nonzero entries they contain, we call them even vectors, and if a vector is not even, we call it odd. As a consequence of the degeneracy of the bilinear form, there are some striking differences with frame theory over real or complex Hilbert spaces. In this paper, the authors explore the construction and properties of binary Parseval frames, and compare them with real and complex ones. Their main results are as follows. In the real or complex case, it is known that each Parseval frame has a Naimark complement. The complementarity is most easily formulated by stating that the Gram matrices of two complementary Parseval frames sum to the identity. The paper shows that in the binary case, not every Parseval frame has a Naimark complement. It also shows that a necessary and sufficient condition for its existence is that the Parseval frame contains at least one even vector. Moreover, the authors study the structure of Gram matrices. The Gram matrices of real or complex Parseval frames are characterized as symmetric or hermitian idempotent matrices. The binary case requires the additional condition that at least one column vector of the matrix is odd.
0 references
frames
0 references
Parseval frames
0 references
binary Parseval frame
0 references
binary cyclic frame
0 references
finite-dimensional vector spaces
0 references
binary numbers
0 references
orthogonal extension principle
0 references
switching equivalence
0 references
Naimark complement
0 references
Gram matrices
0 references
Gram-Schmidt orthogonalization
0 references
0 references
0 references