In this paper, we propose a novel variable-rate learned image compression framework with a conditional autoencoder. Previous learning-based image compression methods mostly require training separate networks for different compression rates so they can yield compressed images of varying quality. In contrast, we train and deploy only one variable-rate image compression network implemented with a conditional autoencoder. We provide two rate control parameters, i.e., the Lagrange multiplier and the quantization bin size, which are given as conditioning variables to the network. Coarse rate adaptation to a target is performed by changing the Lagrange multiplier, while the rate can be further fine-tuned by adjusting the bin size used in quantizing the encoded representation. Our experimental results show that the proposed scheme provides a better rate-distortion trade-off than the traditional variable-rate image compression codecs such as JPEG2000 and BPG. Our model also shows comparable and sometimes better performance than the state-of-the-art learned image compression models that deploy multiple networks trained for varying rates.
. Can. J. Chem. 66,2104Chem. 66, (1988. Absolute oscillator strength spectra in the C l s (280-340 eV) and F l s (680-740 eV) regions of the pefiuoro-n-alkanes from C2 to C6 and perfluorocycloalkanes from C3 to C6 have been determined from inner-shell electron energy loss spectra recorded under electric-dipole scattering conditions. The spectral features are interpreted in terms of spatially localized transitions terminating at orbitals of predominantly a*(C-F) and a*(C-C) character. When compared to the spectra of the perfluoro-n-alkanes, both the C 1s and F 1s spectra of the perfluorocycloalkanes exhibit additional low-lying bands which are assigned to transitions terminating at a*(C-C) orbitals which are shifted to low energy by the combination of the strain of cyclization and the inductive effect of the fluorination. The electron transmission spectra of selected perfluorocycloalkanes (which provide information on their anion states) show as well that the electron affinities of the cyclic systems are substantially lower than those of the corresponding perfluoro-n-alkanes, again as a result of a low-lying a * orbital in the cyclic species. Quantum chemical calculations of the alkane and perfluoroalkane ground-state orbital structures support the experimental results. The localized character of the inner-shell excitations, indicated by the constancy of both term values and oscillator strengths with increasing chain length, contrasts with the more delocalized character of the states accessed in ultraviolet excitation or negative ion formation. De plus, les spectres de transminnion electronique des ions nCgatifs de perfluorocycloalcanes choisis montrent que les affinitits tlectroniques des systttmes cycliques sont beaucoup plus faibles que celles des perfluoro-n-alcanes correspondants; ces rCsultats sont encore une fois expliquCs par la presence d'orbitales a * basses dans les systtmes cycliques. Des calculs de chimie quantique effectuts sur les structures des orbitales des Ctats donsamentaux d'alcanes et de perfluorocycloalcanes sont en accord avec les rtsultats expCrimentaux. Le caracttre localis6 des excitations de couches internes, tel qulindiquC par la constance de chacun des termes des valeurs ainsi que des forces d'oscillation avec une augmentation de la longueur de la chaine, est en opposition avec le caractkre plus dClocalisC des Ctats atteints par les excitations ultraviolettes ou par la formation d'ions negatifs.[Traduit par le revue]
The stabilization method is used to characterize the temporary anion states of C2F6. Based on the theoretical results, we assign the lowest-energy feature in the electron transmission spectrum of C2F6 to the 2Eu ground-state anion and the second feature to overlapping 12Au, 12A1g, 22Au, and 12Eg anion states. It is also proposed that the bimodal kinetic energy distribution of the F- ions produced in the dissociative attachment of near 5 eV electrons on C2F6 is due to the Jahn−Teller effect.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.