RD-Optimized Trit-Plane Coding of Deep Compressed Image Latent Tensors

TL;DR

This paper enhances DPICT, a learning-based image codec, by introducing efficient trit-plane slicing and RD-optimized coding, significantly reducing complexity and improving rate-distortion performance.

Contribution

It proposes a parallel computing scheme for probability estimation and demonstrates that trit-plane slicing outperforms bit-plane slicing in RD performance.

Findings

Reduced time complexity through parallel probability computation

Trit-plane slicing yields better RD performance than bit-plane slicing

Significant efficiency gains in DPICT implementation

Abstract

DPICT is the first learning-based image codec supporting fine granular scalability. In this paper, we describe how to implement two key components of DPICT efficiently: trit-plane slicing and rate-distortion-optimized (RD-optimized) coding. In DPICT, we transform an image into a latent tensor, represent the tensor in ternary digits (trits), and encode the trits in the decreasing order of significance. For entropy encoding, it is necessary to compute the probability of each trit, which demands high time complexity in both the encoder and the decoder. To reduce the complexity, we develop a parallel computing scheme for the probabilities, which is described in detail with pseudo-codes. Moreover, we compare the trit-plane slicing in DPICT with the alternative bit-plane slicing. Experimental results show that the time complexity is reduced significantly by the parallel computing and that the trit-plane slicing provides better RD performances than the bit-plane slicing.