STanH : Parametric Quantization for Variable Rate Learned Image Compression

• Main Authors: Presta, Alberto, Tartaglione, Enzo, Fiandrotti, Attilio, Grangetto, Marco
• Format: Journal Article
• Language: English
• Published: 01-10-2024
• Subjects: Computer Science - Computer Vision and Pattern Recognition; Computer Science - Multimedia
• DOI: 10.48550/arxiv.2410.00557

In end-to-end learned image compression, encoder and decoder are jointly
trained to minimize a $R + {\lambda}D$ cost function, where ${\lambda}$
controls the trade-off between rate of the quantized latent representation and
image quality. Unfortunately, a distinct encoder-decoder pair with millions of
parameters must be trained for each ${\lambda}$, hence the need to switch
encoders and to store multiple encoders and decoders on the user device for
every target rate. This paper proposes to exploit a differentiable quantizer
designed around a parametric sum of hyperbolic tangents, called STanH , that
relaxes the step-wise quantization function. STanH is implemented as a
differentiable activation layer with learnable quantization parameters that can
be plugged into a pre-trained fixed rate model and refined to achieve different
target bitrates. Experimental results show that our method enables variable
rate coding with comparable efficiency to the state-of-the-art, yet with
significant savings in terms of ease of deployment, training time, and storage
costs