# Patch-Based Image Coding With End-to-End Learned Codec Using Overlapping _by Marwa Tarchouli (IETR VAADER) - 2022.11.22_ ###### tags: `VAADER` `Seminar` ![](https://i.imgur.com/IXBbrUX.png) ## Video {%youtube g2F8pm5wpOg%} ## Abstract End-to-end learned image and video codecs, based on auto-encoder architecture, adapt naturally to image resolution, thanks to their convolutional aspect. However, while coding high resolution images, these codecs face hardware problems such as memory saturation. Patch-based image coding with learned image codecs solve these problems. Nevertheless, this method creates block artifacts in the reconstructed image which highly deteriorates its quality. This paper proposes a patch-based image coding solution, which is based on an end-to-end learned model, and which remedies the block artifacts. It consists in coding overlapping patches of the image and reconstruct them into a decoded image using a weighting function. This approach manages to slightly outperform the performance of full résolution image coding with an end-to-end learned model while being adaptable to different memory size.