no code implementations • 12 Aug 2023 • Kristian Fischer, Christian Herglotz, André Kaup
Coding 4K data has become of vital interest in recent years, since the amount of 4K data is significantly increasing.
no code implementations • 12 Jul 2023 • Fabian Brand, Alexander Kopte, Kristian Fischer, André Kaup
Current state-of-the-art neural-network-based image compression systems however use only one scale to transmit the latent space.
1 code implementation • 25 Apr 2023 • Christian Herglotz, Hannah Och, Anna Meyer, Geetha Ramasubbu, Lena Eichermüller, Matthias Kränzler, Fabian Brand, Kristian Fischer, Dat Thanh Nguyen, Andy Regensky, André Kaup
Using additional supporting points inbetween standard points defined by parameters such as the quantization parameter, we assess the interpolation error of the Bj{\o}ntegaard-Delta (BD) calculus and its impact on the final BD value.
no code implementations • 27 Feb 2023 • Kristian Fischer, Fabian Brand, Christian Blum, André Kaup
Areas without saliency are transmitted in latent spaces of lower spatial resolution in order to reduce the bitrate.
1 code implementation • 20 Jan 2023 • Kristian Fischer, Fabian Brand, Christian Herglotz, André Kaup
This is a well-known method for the human visual system, where scaling lists are derived from psycho-visual models.
no code implementations • 13 May 2022 • Kristian Fischer, Markus Hofbauer, Christopher Kuhn, Eckehard Steinbach, André Kaup
To mitigate this problem, we propose an evaluation method based on pseudo ground-truth data from the field of semantic segmentation to the evaluation of video coding for machines.
no code implementations • 13 May 2022 • Kristian Fischer, Christian Forsch, Christian Herglotz, André Kaup
Video and image coding for machines (VCM) is an emerging field that aims to develop compression methods resulting in optimal bitstreams when the decoded frames are analyzed by a neural network.
no code implementations • 13 May 2022 • Kristian Fischer, Christian Blum, Christian Herglotz, André Kaup
Deep neural object detection or segmentation networks are commonly trained with pristine, uncompressed data.
no code implementations • 11 Mar 2022 • Kristian Fischer, Christian Herglotz, André Kaup
Classical video coding for satisfying humans as the final user is a widely investigated field of studies for visual content, and common video codecs are all optimized for the human visual system (HVS).
no code implementations • 11 Mar 2022 • Kristian Fischer, Felix Fleckenstein, Christian Herglotz, André Kaup
Saliency-driven image and video coding for humans has gained importance in the recent past.
no code implementations • 11 Mar 2022 • Kristian Fischer, Fabian Brand, Christian Herglotz, André Kaup
In this paper, we propose a standard-compliant feature-based RDO (FRDO) that is designed to increase the coding performance, when the decoded frame is analyzed by a neural network in a video coding for machine scenario.
no code implementations • 1 Mar 2022 • Simon Grosche, Kristian Fischer, Fabian Brand, Jürgen Seiler, André Kaup
Quarter sampling is a novel sensor concept that enables the acquisition of higher resolution images without increasing the number of pixels.
no code implementations • 5 Jan 2022 • Fabian Brand, Kristian Fischer, Alexander Kopte, André Kaup
Even though rate-distortion optimization is a crucial part of traditional image and video compression, not many approaches exist which transfer this concept to end-to-end-trained image compression.
1 code implementation • 15 Dec 2021 • Kristian Fischer, Fabian Brand, André Kaup
Compared to the standard VVC, 41. 4% of bitrate are saved by this method for Mask R-CNN as analysis network on the uncompressed Cityscapes dataset.