iWave3D: End-to-end Brain Image Compression with Trainable 3-D Wavelet Transform

With the rapid development of whole brain imaging technology, a large number of brain images have been produced, which puts forward a great demand for efficient brain image compression methods. At present, the most commonly used compression methods are all based on 3-D wavelet transform, such as JP3D. However, traditional 3-D wavelet transforms are designed manually with certain assumptions on the signal, but brain images are not as ideal as assumed. What's more, they are not directly optimized for compression task. In order to solve these problems, we propose a trainable 3-D wavelet transform based on the lifting scheme, in which the predict and update steps are replaced by 3-D convolutional neural networks. Then the proposed transform is embedded into an end-to-end compression scheme called iWave3D, which is trained with a large amount of brain images to directly minimize the rate-distortion loss. Experimental results demonstrate that our method outperforms JP3D significantly by 2.012 dB in terms of average BD-PSNR.

I. Introduction

In recent years, the rapid development of brain imaging technology has promoted the advancement of many fields such as medicine and artificial intelligence. However, as the center of the nervous system, the brain usually has a complex structure, and imaging it will generate massive amounts of data. For example, the brain of drosophila contains about 10⁵ of neurons, and its raw images occupy about 106 terabyte (TB) of storage space [1]. The human brain has more than 1.5×10¹⁰ neurons, and the space required to store its raw images will be unimaginable. This puts forward a great demand for efficient brain image compression methods.