International Journal of Intelligent Systems Technologies and Applications 

About: International Journal of Intelligent Systems Technologies and Applications is an academic journal published by Inderscience Publishers. The journal publishes majorly in the area(s): Computer science & Artificial neural network. It has an ISSN identifier of 1740-8865. Over the lifetime, 438 publications have been published receiving 3015 citations. The journal is also known as: IJISTA & Intelligent systems technologies and applications.

Gesture recognition with a Time-Of-Flight camera

Abstract: This paper presents a new approach for gesture classification using x- and y- projections of the image and optional depth features. The system uses a 3D Time-Of-Flight (TOF) sensor which has the big advantage of simplifying hand segmentation. For the presented system, a Photonic-Mixer-Device (PMD) camera with a resolution of 160 × 120 pixels and a frame rate of 15 frames per second is used. The goal of our system is to recognise 12 different static hand gestures. The x- and y- projections and the depth features of the captured image are good enough to use a simple nearest neighbour classifier, resulting in a fast classification. To evaluate the system, a set of 408 images is recorded, 12 gestures from 34 persons. With a 'Leave-One-Out' evaluation, the recognition rate of the system is 94.61% and the classification time is about 30 msec on a standard PC.

Fusion of stereo vision and Time-Of-Flight imaging for improved 3D estimation

Abstract: This paper suggests an approach in fusing two 3D estimation techniques: stereo vision and Time-Of-Flight (TOF) imaging. By mapping the TOF-depth measurements to stereo disparities, the correspondence between the images from a fast TOF-camera and standard high resolution camera pair are found, so the TOF depth measurements can be linked to the image pairs. In the same framework, a method is developed to initialise and constrain a hierarchical stereo matching algorithm. It is shown that in this way, higher spatial resolution is obtained than by only using the TOF camera and higher quality dense stereo disparity maps are the results of this data fusion.

Stream processing for fast and efficient rotated Haar-like features using rotated integral images

Abstract: An extended set of Haar-like features for image sensors beyond the standard vertically and horizontally aligned Haar-like features and the 45° twisted Haar-like features are introduced. The extended rotated Haar-like features are based on the standard Haar-like features that have been rotated based on whole integer pixel-based rotations. These rotated feature values can also be calculated using rotated integral images which mean that they can be fast and efficiently calculated with just eight operations irrespective of the feature size. The integral image calculations can be offloaded to the graphical processing unit (GPU) using the stream processing paradigm. The integral image calculation on the GPU is seen to be faster than the traditional central processing unit implementation of the algorithm, for large image sizes, allowing more complex classifiers to be implemented in real-time.

Pose estimation and map building with a Time-Of-Flight-camera for robot navigation

Abstract: In this paper, we describe a joint approach for robot navigation with collision avoidance, pose estimation and map building with a 2.5D Photonic Mixer Device (PMD)-camera combined with a high-resolution spherical camera. The cameras are mounted at the front of the robot with a certain inclination angle. The navigation and map building consists of two steps: when entering new terrain the robot first scans the surrounding. Simultaneously a 3D-panorama is generated from the PMD-images. In the second step, the robot moves along the pre-defined path, using the PMD-camera for collision avoidance and a combined Structure-from-Motion (SfM) and model-tracking approach for self-localisation. The computed poses of the robot are simultaneously used for map building with new measurements from the PMD-camera.

On-the-fly scene acquisition with a handy multi-sensor system

Abstract: We present a scene acquisition system which allows for fast and simple acquisition of arbitrarily large 3D environments. We propose a small device which acquires and processes frames consisting of depth and colour information at interactive rates. This allows the operator to control the acquisition process on the fly. However, no user input or prior knowledge of the scene is required. In each step of the processing, pipeline colour and depth data are used in combination in order to gain from different strengths of the sensors. A novel registration method is introduced that combines geometry and colour information for enhanced robustness and precision. We evaluate the performance of the system and present results from acquisition in different environments.