| 1. |
|
|
| 2. |
- Barnard, Yvonne, et al.
(författare)
-
Data management and data sharing in field operational tests
- 2016
-
Ingår i: Intelligent Transportation Systems: From Good Practices to Standards. - : CRC Press. - 9781498721875 ; , s. 59-72
-
Bokkapitel (övrigt vetenskapligt/konstnärligt)abstract
- In this chapter it will be discussed how data from Field Operational Tests of Intelligent Transport Systems can be managed and shared. The Field Operational Tests, where hundreds of users get to experience the latest systems, aim to assess the impacts that would result from a wide-scale implementation. Evaluation principles of Field Operational Tests will be explained, and a closer look will be taken at the data that is collected for carrying out the assessments. The widely used FESTA methodology for designing and conducting Field Operational Tests and Naturalistic Driving Studies already provides several recommendations for managing data. This methodology will be discussed and illustrated by examples of its use in European projects. As field test projects set out to collect a huge set of data, the projects themselves do not usually have the scope or the resources to analyze the data from every perspective. Therefore re-use of the collected data also by other projects with different research questions has the potential to generate a wealth of new knowledge about what is happening in the interactions between drivers, vehicles and the infrastructure. Data sharing is the focus of a European support action, FOT-Net Data. The support action is working, with international collaboration, to form a data sharing framework, a data catalogue, and provide detailed recommendations for sharing and re-use. Outcomes from this activity will be discussed. Ways of sharing different types of data will be described, including the necessary steps to be taken to open up the data.
|
|
| 3. |
- Shao, Wen-Ze, et al.
(författare)
-
On potentials of regularized Wasserstein generative adversarial networks for realistic hallucination of tiny faces
- 2019
-
Ingår i: Neurocomputing. - : ELSEVIER. - 0925-2312 .- 1872-8286. ; 364, s. 1-15
-
Tidskriftsartikel (refereegranskat)abstract
- Super-resolution of facial images, a.k.a. face hallucination, has been intensively studied in the past decades due to the increasingly emerging analysis demands in video surveillance, e.g., face detection, verification, identification. However, the actual performance of most previous hallucination approaches will drop dramatically when a very low-res tiny face is provided, due to the challenging multimodality of the problem as well as lack of an informative prior as a strong semantic guidance. Inspired by the latest progress in deep unsupervised learning, this paper focuses on tiny faces of size 16 x 16 pixels, hallucinating them to their 8 x upsampling versions by exploring the potentials of Wasserstein generative adversarial networks (WGAN). Besides a pixel-wise L2 regularization term imposed to the generative model, it is found that our advocated autoencoding generator with both residual and skip connections is a critical component for WGAN representing the facial contour and semantic content to a reasonable precision. With the additional Lipschitz penalty and architectural considerations for the critic in WGAN, the proposed approach finally achieves state-of-the-art hallucination performance in terms of both visual perception and objective assessment. The cropped CelebA face dataset is primarily used to aid the tuning and analysis of the new method, termed as tfh-WGAN. Experimental results demonstrate that the proposed approach not only achieves realistic hallucination of tiny faces, but also adapts to pose, expression, illuminance and occluded variations to a great degree.
|
|
