Keyless entry systems in cars are adopting neural networks for localizing its operators. Using test-time adversarial defences equip such systems with the ability to defend against adversarial attacks without prior tra...
详细信息
Structure-from-Motion (SfM) approaches, which are conventionally based on local interest point matches, tend to work well for richly textured indoor- and outdoor environments. However, in less textured scene areas the...
This paper proposes an efficient approach for semantic image classification by integrating additional contextual constraints such as class co-occurrences into a randomized forest classification framework. The randomiz...
详细信息
The tasks carried out by modern information workers become increasingly complex and time-consuming. They often require to evaluate, interpret, and compare information from different sources presented in multiple appli...
详细信息
ISBN:
(纸本)9781568817125
The tasks carried out by modern information workers become increasingly complex and time-consuming. They often require to evaluate, interpret, and compare information from different sources presented in multiple application windows. With large, high resolution displays, multiple application windows can be arranged in a way so that a large amount of information is visible simultaneously. However, individual application windows' contents and visual representations are isolated and relations between information items contained in these windows are not explicit. Thus, relating and comparing information across applications has to be executed manually by the user, which is a tedious and error-prone task. In this paper we present visual links connecting related pieces of information across application windows and thereby guiding the user's attention to relevant information. Applications are coordinated by a management application accessible via a light-weight interface. User selections are synchronized across registered applications and visual links are rendered on top of the desktop content by a window manager. Initial user feedback was very positive and indicates that visual links improve task efficiency when analyzing information from multiple sources.
Traditional Structure-from-Motion (SfM) approaches work well for richly textured scenes with a high number of distinctive feature points. Since man-made environments often contain textureless objects, the resulting po...
详细信息
computervision-based interfaces to games hold the promise of rich natural interaction and thus a more realistic gaming experience. Therefore, the video games industry started to develop and market computervision-bas...
详细信息
Tracking and detection of objects often require to apply complex models to cope with the large intra-class variability of the foreground as well as the background class. In this work, we reduce the complexity of a bin...
详细信息
Object detection and recognition algorithms are an integral part of the architecture of many modern image processing systems employing computervision (CV) techniques. In this paper we describe our work in the area of...
详细信息
In this paper we propose an efficient unsupervised texture segmentation method. We introduce a texture extension of a state-of-the-art color segmentation algorithm. We show how to use covariance matrices of low level ...
详细信息
The usage of projected displays in everyday office environments is still uncommon due to deficiencies in nowadays projection systems. An automatic calibration routine for multi-projector setups should adopt to the exi...
详细信息
暂无评论