Abstract
Recently, convolutional neural networks have been successfully applied to lossy image compression. End-to-end optimized autoencoders, possibly variational, are able to dramatically outperform traditional transform coding schemes in terms of rate-distortion trade-off; however, this is at the cost of a higher computational complexity. An intensive training step on huge databases allows autoencoders to learn jointly the image representation and its probability distribution, possibly using a non-parametric density model or a hyperprior auxiliary autoencoder to eliminate the need for prior knowledge. However, in the context of on board satellite compression, time and memory complexities are submitted to strong constraints. The aim of this paper is to design a complexity-reduced variational autoencoder in order to meet these constraints while maintaining the performance. Apart from a network dimension reduction that systematically targets each parameter of the analysis and synthesis transforms, we propose a simplified entropy model that preserves the adaptability to the input image. Indeed, a statistical analysis performed on satellite images shows that the Laplacian distribution fits most features of their representation. A complex non parametric distribution fitting or a cumbersome hyperprior auxiliary autoencoder can thus be replaced by a simple parametric estimation. The proposed complexity-reduced autoencoder outperforms the Consultative Committee for Space Data Systems standard (CCSDS 122.0-B) while maintaining a competitive performance, in terms of rate-distortion trade-off, in comparison with the state-of-the-art learned image compression schemes.
Highlights
Satellite imaging has many applications in oceanography, agriculture, biodiversity conservation, forestry, landscape monitoring, geology, cartography or military surveillance [1].The increasing spectral and spatial resolutions of on board sensors allow obtaining everbetter quality products, at the cost of an increased amount of data to be handled
In addition to the mean square error (MSE), we evaluate those results in terms of multi-scale structural similarity index (MS-SSIM)
For either low or high bit rates, a drastic reduction of N starting from the reference architecture [16], does not decrease significantly the performance, both in MSE and in MS-SSIM, while it leads to a complexity decrease of more than 70%
Summary
The increasing spectral and spatial resolutions of on board sensors allow obtaining everbetter quality products, at the cost of an increased amount of data to be handled. In this context, on board compression plays a key role to save transmission channel bandwidth and to reduce data-transmission time [2]. Lossless compression is a reversible technique that compresses data without loss of information. In the context of image compression, autoencoders are used to learn a representation with low entropy after quantization. The autoencoder is composed of an analysis transform and a synthesis transform connected by a bottleneck that performs quantization and coding. We focus on two reference architectures: [13] displayed in Figure 1 (left) and [16]
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
