Fig 1 - uploaded by Dirk Hölscher
Content may be subject to copyright.
![Chest CT Example [30]](profile/Dirk-Hoelscher-2/publication/376314798/figure/fig1/AS:11431281234075431@1712221184475/Chest-CT-Example-30.jpg)
Source publication
![Fig. 1. Chest CT Example [30]](profile/Dirk-Hoelscher-2/publication/376314798/figure/fig1/AS:11431281234075431@1712221184475/Chest-CT-Example-30_Q320.jpg)
Hyperparameter tuning is an important aspect in machine-learning especially for deep generative models. Tuning models to stabilize training and to get the best accuracy can be a time consuming and protracted process. Generative models have a large search space requiring resources and knowledge to find the best parameters. Therefore, in most cases t...
Context in source publication
Context 1
... dataset used to get these initial values (i.e., Pix2Pix tuning) is the Chest CT [30] dataset from Kaggle. Figure 1 shows an example image of the dataset used to calculate the UIQ scores for the dataset used in this paper. The baseline UIQ was received by training a Pix2Pix network with standard parameters to create a copy of the input and calculating similarity. ...
Similar publications
We provide the theoretical foundation for the recently proposed tests of equal forecast accuracy and encompassing by Pitarakis (2023a) and Pitarakis (2023b), when the competing forecast specification is that of a factor-augmented regression model, whose loadings are allowed to be homogeneously/heterogeneously weak. This should be of interest for pr...
Citations
... This data was used to train a Pix2Pix GAN which was trained to create exact copies [10] of the scans. Based on our previous work [11], [12] and [13] we optimised Pix2Pix to create high quality samples using the evaluation of the Universal Quality Index Metric (UIQ) [14] to optimise the generated images towards an ideal UIQ score by using hyperparameter tuning and then evaluating the results using UIQ. In addition, we created, based on our experiments a prediction network which is able to predict if a hyperparameter combination is able to generate better results or not. ...
... Generative models attracted researchers following high feature space generation performance and good predictions. Losses structure of 2D images in pix2pix GAN problems are re-dominantly formulated as per 2D image pixel classification rate or regression formulated as per 2D image pixel classification rate or regression [41]- [44]. These formulations present the unstructured output area in the orientation of output pixels, which means each of them is independent of all others in input 2D images. ...
Research on training Generative Adversarial Networks (GANs) to create 3D human body avatars from 2D datasets is underway. Research done in this field shows promise and has paved the way for significant advancements in a variety of applications, including virtual reality, sports analysis, cinematography, surveillance, and cinematography. By avoiding obstacles and producing high-resolution, rich-information multi-view (RGB) images, drone active tracking combined with aerial photography sensors can eliminate occlusions and enable 3D avatar body reconstruction. Due to several issues, such as restricted perspective coverage, obvious occlusions, and texture disappearance, 3D avatar reconstruction techniques encounter training failures that cause distortions and feature loss in 3D reconstructed models. The new end-to-end trainable deep neural network methodology PIXGAN-Drone is presented in this paper for the photo-realistic 3D avatar of the human body reconstruction from multi-view images. It is based on the integration of active tracking drones equipped with aerial photography sensors (stable automatic circular motion system) into the Pix2Pix GANs training framework to generate high-resolution 2D models. Conditional image-to-image translation and dynamic aerial perspectives can be used to develop realistic and accurate 3D models. This research conducted experiments on multiple datasets to demonstrate the improved performance of our method over state-of-the-art methods for various metrics (Chamfer, P2S, and CED). The results showed that our 3D reconstructed human avatars were 0.0293, 0.0271, and 0.0232 on RenderPeople, 0.0133, 0.0136, 0.0050 on People Snapshot (indoor), 0.0154, 0.0101, 0.0063 on People Snapshot (outdoor), and 0.0316, 0.0275, 0.0216 on Custom data-drone (collected dataset).
