Fachbereich Informatik und Medien
Refine
Year of publication
Document Type
- Conference Proceeding (294)
- Article (100)
- Part of a Book (28)
- Lecture (21)
- Bachelor Thesis (14)
- Book (12)
- Master's Thesis (11)
- Report (6)
- Moving Images (3)
- Other (3)
Keywords
- Compositing (9)
- Digitalkamera (9)
- Farbbildverarbeitung (9)
- HDTV (9)
- Kameraarbeit (9)
- Videobearbeitung (9)
- cybersecurity (9)
- Dreidimensionale Video (8)
- Filmproduktion (8)
- Shodan (7)
Institute
With the release of the Apple iPhone 12 Pro Max in 2020, various features were integrated that make it attractive as a recording device for scene-related computer graphics pipelines. The captured Apple RAW images have a much higher dynamic range than the standard 8-bit images. Since a scene-based workflow naturally has an extended dynamic range (HDR), the Apple RAW recordings can be well integrated. To correctly integrate the iPhone 12 pro Apple RAW data into a scene-related workflow, two command-line software solutions, among others, are dcraw and rawtoaces. In this paper, we concentrate on dcraw. dcraw offers the possibility to export RAW images directly to ACES2065-1. Different images recorded under different lighting conditions are examined to determine which options for the dcraw development produce the best results. Images from five different lighting situations like D65, D60, D55, D50, and 2980K, are examined. The DNG files are developed using different methods within dcraw. The values of the recorded ColorChecker24 patches are compared to reference values derived from BabelColor [1] and ACES [2].
In this work, the most relevant 3D LiDAR technologies and their applications in 2022 were investigated. For this purpose, applications of LiDAR systems were classified into the typical application areas "3D modeling", "smart city", "robotics", "smart automotive" and "consumer goods". The investigation has shown that neither "mechanical" LiDAR technologies, nor so-called solid-state LiDAR technologies, nor "hybrid" LiDAR technologies can be evaluated as optimal for the typical application areas. In none of the application areas could all of the elaborated requirements be met. However, the "hybrid" LiDAR technologies such as sequential MEMS LiDAR technology and sequential flash LiDAR technology proved to be among the most suitable for most typical application areas. However, other technologies also tended to be suitable for individual typical application areas. Finally, it was found that several of the LiDAR technologies investigated are currently equally suitable for some typical application areas. To evaluate the suitability, concrete LiDAR systems - of different technologies and properties - were compared with the specific requirements of exemplary applications of an application area. The results of the investigation provide an orientation as to which LiDAR technology is promising for which application area.
This Paper proposes a novel method for the blind detection of image pre-processing techniques by means of statistical pattern recognition in image forensics. The technique is intended to detect sensor intrinsic pre-processing steps as well as manually applied filters. We have exemplary chosen 6 pre-processing filters with different parameter settings. The concept utilizes 29 image features which are supposed to allow for a reliable model creation during supervised learning. The evaluation of the trained models indicates average accuracies between 82.50 and 94.53%. The investigation of image data from 8 sensors leads to the detection of credible pre-processing filters. Those results adumbrate that our method might be suitable to prove the authenticity of the data origin and the integrity of image data based on the detected preprocessing techniques. The preliminary evaluation for manually applied filters yields recognition accuracies between 39.09% (14 classes) and 53.33% (7 classes).
The importance and the attractivity of the sensorimotor training is still growing. Up to now the impact of the training on the body is not yet fully investigated. Hence, nowadays the planning of the therapy is mainly based on the experiences of the physiotherapist and on the conditions of the practice. For the development of the therapy the physiotherapist is supported by manufacturer's information about the exercisers as well as by general assumptions regarding the sensorimotor training. For the validation of this information two setups were investigated. In the first part, the behavior of two students was studied on three exercisers. Here the EMG data and the motion data were analyzed. In the second part, the behavior of the left and the right body side was analyzed for 16 subjects. The study revealed that the major work for the maintenance of the equilibrium is done by the distal musculature. Furthermore, it was shown that there is a different behavior of the musculature at both body sides. Additionally, it has been proven that each test person had an individual behavior on the exercisers. Consequently, it would be hard to make general assumptions regarding the impact of the training on the body.
Surface classification and detection of latent fingerprints based on 3D surface texture parameters
(2012)
Since the face morphing attack was introduced by Ferrara et al. in 2014, the detection of face morphings has become a wide spread topic in image forensics. By now, the community is very active and has reported diverse detection approaches. So far, the evaluations are mostly performed on images without post-processing. Face images stored within electronic machine readable documents (eMRTD) are ICAO-passport-scaled to a resolution of 413x531 and a JPG or JP2 lesize of 15 kilobytes. This paper introduces a face morphing detection concept with 3 modules (ICAO-aligned pre- processing module, feature extraction module and classi cation module), tailored for such images on eMRTD. In this work we exemplary design and evaluate two feature spaces for the feature extraction module, a frequency domain and a spatial domain feature space. Our evaluation will compare both feature spaces and is carried out with 66,229 passport-scaled images (64,363 morphed face images and 1,866 authentic face images) which are completly independent from training and include all images provided for the IHMMSEC'19 special session: "Media Forensics - Fake or Real?". Furthermore, we investigate the in uence of di erent morph gen- eration pipelines to the detection accuracies of the concept and we analyse the impact of neutral and smiling genuine faces to the morph detector performance. The evaluation determines a detection rate of 86.0% for passport-scaled morphed images with a false alarm rate of 4.4% for genuine images for the spatial domain feature space.