We present PLONQ, a progressive neural image compression scheme which pushes the boundary of variable bitrate compression by allowing quality scalable coding with a single bitstream. In contrast to existing learned va...
详细信息
ISBN:
(纸本)9781665441155
We present PLONQ, a progressive neural image compression scheme which pushes the boundary of variable bitrate compression by allowing quality scalable coding with a single bitstream. In contrast to existing learned variable bitrate solutions which produce separate bitstreams for each quality, it enables easier rate-control and requires less storage. Leveraging the latent scaling based variable bitrate solution, we introduce nested quantization, a method that defines multiple quantization levels with nested quantization grids, and progressively refines all latents from the coarsest to the finest quantization level. To achieve finer progressiveness in between any two quantization levels, latent elements are incrementally refined with an importance ordering defined in the rate-distortion sense. To the best of our knowledge, PLONQ is the first learning-based progressive image coding scheme and it outperforms SPIHT, a well-known wavelet-based progressive image codec.
We investigate image and video transmission in camera-based wireless sensor networks (WSNs). Our objective is to design a mechanism that finds the transmission strategies that maximize the expected reconstruction qual...
详细信息
ISBN:
(纸本)9781424413539
We investigate image and video transmission in camera-based wireless sensor networks (WSNs). Our objective is to design a mechanism that finds the transmission strategies that maximize the expected reconstruction quality of the transmitted data at the base station while making efficient use of the limited resources of the WSN. Our approach is based on quality-scalablecoding and multipath transmission. Unlike conventional approaches where data is sent over the paths regardless of its importance at the destination, we propose a network-adaptive transmission mechanism that decomposes the source bitstream into segments of unequal importance and reserves the most reliable paths to transmit the segments with the highest importance. Moreover, we use unequal segment loss protection with erasure codes of different strengths to maximize the expected quality at the destination and propose a fast algorithm that finds nearly optimal transmission strategies. We also present simulation results using a test video sequence that show that the proposed transmission mechanism provides adaptability to the time and space varying conditions of camera-based wireless sensor networks.
In this paper, a quality scalable coding for a selected region is presented. If one region is semantically more important than others, it is necessary for an image compression scheme to be capable of handling the regi...
详细信息
In this paper, a quality scalable coding for a selected region is presented. If one region is semantically more important than others, it is necessary for an image compression scheme to be capable of handling the regional semantic difference, because the information loss in the region of interest is more serious. We propose a quality scalable coding along with its model by introducing a quality scale parameter. It is a more extended and generalized image compression philosophy than the conventional coding. For an implementation of the proposed quality scalable coding, an H.263-based scheme is presented. This scheme controls temporal and spatial quality efficiently and improves the reconstructed image quality of the region of interest. (C) 1999 Elsevier Science B.V. All rights reserved.
暂无评论