Optimising Perceived Distortion in Lossy Encoding of Dynamic Meshes

Abstract

Development of geometry data compression techniques in the past years has been limited by the lack of a metric with proven correlation with human perception of mesh distortion. Many algorithms have been proposed, but usually the aim has been to minimise mean squared error, or some of its derivatives. In the field of dynamic mesh compression, the situation has changed with the recent proposal of the STED metric, which has been shown to capture the human perception of mesh distortion much better than previous metrics. In this paper we show how existing algorithms can be steered to provide optimal results with respect to this metric, and we propose a novel dynamic mesh compression algorithm, based on trajectory space PCA and Laplacian coordinates, specifically designed to minimise the newly proposed STED error. Our experiments show that using the proposed algorithm, we were able to reduce the required data rate by up to 50% while preserving the introduced STED error.

Compressing dynamic meshes with geometric laplacians

“…While the algorithm is based on similar ideas as [VS07] and [VP11], a significant gain in the compression performance is due to both the proposed application of the geometric Laplacian and the mesh averaging technique. The geometric Laplacian substantially reduces the redundancy in differential coordinates over the sequence.…”

“…The geometric Laplacian substantially reduces the redundancy in differential coordinates over the sequence. However, in order to build the geometric Laplacian, our technique requires to explicitly encode the average mesh, while [VP11] does not require to store any additional data, since it employs a purely combinatorial Laplacian that can be recovered from connectivity only. A further improvement in the compression performance is achieved by means of the proposed differential coordinates predictor.…”

“…A further improvement in the compression performance is achieved by means of the proposed differential coordinates predictor. The predictor, while similar to other general techniques, is also novel in terms of formulation and application to the framework, and can potentially be applied to similar approaches, such as [VP11], to improve their performance.…”

“…Each vertex trajectory is expressed as a combination of the basis trajectories and the algorithm encodes only the basis trajectory vectors and the coefficients. The same approach has been revised and improved in [VS09], proposing an alternative approach to compress the basis vectors named CO-BRA, and in [VP11], where the CODDYAC algorithm is combined with the Laplacian encoding technique by Sorkine et al…”

“…The rationale behind this approach is that the average mesh contains the sampling information shared by all the other frames. For the data sets that we consider, we achieve a data rate reduction of 10-40%, compared to [VP11], keeping the same amount of distortion as measured by the STED error metric [VS11].…”