Task-Oriented Multi-Bitstream Optimization for Image Compression and Transmission via Optimal Transport

Sa Yan, Nuowen Kan, Chenglin Li, Wenrui Dai, Junni Zou, Hongkai Xiong

Abstract:

Image compression for machine vision exhibits various rate-accuracy performance across different downstream tasks and content types.
An efficient utilization of constrained network resource for achieving an optimal overall task performance has thus recently attracted a growing attention. In this paper, we propose Tombo, a task-oriented image compression and transmission framework that efficiently identifies the optimal encoding bitrate and routing scheme for multiple image bitstreams delivered simultaneously for different downstream tasks. Specifically, we study the characteristics of image rate-accuracy performance for different machine vision tasks, and formulate the task-oriented joint bitrate and routing optimization problem for multi-bitstreams as a multi-commodity network flow problem with the time-expanded network modeling. To ensure consistency between the encoding bitrate and routing optimization, we also propose an augmented network that incorporates the encoding bitrate variables into the routing variables. To improve computational efficiency, we further convert the original optimization problem to a multi-marginal optimal transport problem, and adopt a Sinkhorn iteration-based algorithm to quickly obtain the near-optimal solution. Finally, we adapt Tombo to efficiently deal with the dynamic network scenario where link capacities may fluctuate over time. Empirical evaluations on three typical machine vision tasks and four real-world network topologies demonstrate that Tombo achieves a comparable performance to the optimal one solved by the off-the-shelf solver Gurobi, with a $5\times \sim 114\times$ speedup.

Primary Subject Area: [Systems] Transport and Delivery

Relevance To Conference:

The problem addressed in this paper involves optimizing the image compression and delivery over the communication networks for implementing multiple machine vision tasks, which belongs to the field of image processing and communications, a major topic of multimedia processing and communications. Specifically, we establish a task-oriented bitrate allocation and routing optimization problem for the joint encoding and delivery of multiple image bitstreams for better implementing multiple machine vision tasks, such as the image segmentation and classification. The objective is to improve the task performance and reduce the delivery costs under limited network delivery capacities. Our proposed method improves the computational efficiency of providing a near-optimal solution to the joint image compression and delivery optimization problem with a 5$\times \sim 144 \times$ speedup compared to the baseline algorithms, which facilitates a fine-grained performance optimization for online machine vision tasks under time-varying network conditions.

In addition, the conference of ACM MM has published numerous works related to the image/video transport and delivery, indicating the relevance of our paper to the field of multimedia processing.

Submission Number: 4489