Recently, data driven algorithms such as neural networks have attracted a lot of attention and have become a popular area of research and development. This interest is driven by several factors, such as recent advances in processing power (cheap and powerful hardware), the availability of large data sets (big data) and several small but important algorithmic advances. Nowadays, neural networks are the state of the art for several computer vision tasks, such as the ones that require high level understanding of image semantics, e.g. image classification, object segmentation, saliency detection, but also in low level image processing tasks, such as image denoising, inpainting, super-resolution and so on. These advances have led to an increased interest in applying deep neural networks to the problem of image compression, which is the main focus of this project. More precisely, the aim is to find a compact image representation model that has been obtained (learned) with a huge amount of visual data and can represent the wide variety of visual content that is available today in an efficient way. Some of the early solutions in the learning-based image coding field already show encouraging results in terms of rate-distortion performance, especially in comparison with conventional image codecs (e.g. JPEG 2000 and HEVC Intra) that compress the visual information with hand-crafted transforms, entropy coding and quantization schemes.
|Start Date: 01-06-2021|
|End Date: 01-06-2024|
|Team: João Miguel Duarte Ascenso, Catarina Isabel Carvalheiro Brites Ascenso, Fernando Manuel Bernardo Pereira|
|Groups: Multimedia Signal Processing – Lx|
|Local Coordinator: João Miguel Duarte Ascenso|