Paper 2011/631 has a fundamental flaw
Posted by: robdep
Date: 03 December 2011 15:17
This paper attacks the results of the De Prisco and De Santis paper (hereafter referred to as DD) using a model that is different from the model assumed in the DD paper. Hence this paper is meaningless.
In the original model of the DD paper the attack described in this paper is not possible.
The authors of this paper can obviously point out that the scheme presented in th DD paper are not immune to cheating in the model they consider, but this point has been already noted in the DD paper.
Probably the mistake is due to the fact that the authors of this paper have not noticed that the DD paper uses a model that is slightly stronger then the usual one. The extra restriction assumed in the DD model were assumed just to avoid the kind of attack the the authors present in this paper. And the proofs and explanations in
the DD paper make this point clear enough:
1. Definition 3.1 of the model, on page 3, second column;
2. Sentence on page 9, first column, lines 20-25;
3. Sentence on page 9, second column, lines 24-26;
4. Argumentation for the proof of Lemma 6.3
(on page 10, second column, Case 4).
In particular the above cited argumentation explicitly says that it is possible for the cheaters to construct a fake share that superposed to the one of the honest participant will yield a number of black subpixels STRICTLY GREATER than the threshold h, but the honest participant CAN DETECT this situation because legitimate shares produce a reconstructed black pixel with EXACTLY h black subpixels, as required by the model used in the DD paper.
Edited 1 time(s). Last edit at 03-Dec-2011 15:33 by robdep.