China
Africa
Explore chapters and articles related to this topic
Principles for successful email
Published in Gerald Rau, Writing for Engineering and Science Students, 2019
One of the advantages of email is that it provides almost instantaneous correspondence anywhere in the world. The related disadvantage is that we have grown accustomed to an almost instantaneous response. If you need more time to think or act, send a quick acknowledgment and say when you will be able to provide a more complete answer.
RAFDivider: a distributed algorithm for computing semantics in higher-order abstract argumentation frameworks
Published in Journal of Applied Non-Classical Logics, 2023
Sylvie Doutre, Marie-Christine Lagasquie-Schiex
Moreover, considering a formal representation of this reasoning model, argumentation frameworks with higher-order attacks (e.g. Baroni, Cerutti, Dunne et al., 2011; Baroni, Cerutti, Giacomin et al., 2011; Barringer et al., 2005; Modgil, 2007, 2009) are a rich extension of the classical Argumentation Framework (AF) by Dung (1995): not only they consider arguments and attacks between arguments, but also attacks on attacks (see for instance Baroni, Cerutti, Dunne et al., 2011; Baroni, Cerutti, Giacomin et al., 2011). Among these frameworks, the Recursive Argumentation Framework (RAF) by Cayrol et al. (2020) proposes a direct approach regarding acceptability, which outputs sets of arguments and/or attacks (defined under the notion of structure), keeping the full expressiveness of higher-order attacks. A correspondence between Dung's extension-based semantics for AFs and structure-based semantics of RAFs without any attack on attacks has been shown in Cayrol et al. (2020), proving that RAFs are a conservative generalisation of AFs. This characteristic makes RAFs particularly interesting to consider.
Automatic 3D human body landmarks extraction and measurement based on mean curvature skeleton for tailoring
Published in The Journal of The Textile Institute, 2022
Haoyang Xie, Yueqi Zhong, Zhicai Yu, Azmat Hussain, Guanmin Chen
We use the right-hand coordinate system in this paper, and the human body is facing the +Z direction. To alleviate the complexity of the problem, we give two natural assumptions: 1) all human models are standing, and 2) there are no severe distortions in the body, such as limbs crossing, severe torso bending, etc. With these assumptions, we can explicitly establish the correspondence between clustering results and semantics, as shown in Figure 1. The abbreviation for each part in Figure 1 will be used in the rest of this paper. A specific part is denoted as with a subscript, for example. In addition, we also propose a simple and effective method to identify whether the arm is higher or lower than the shoulder, which is necessary for the automatic extraction of specific landmarks. Specifically, the proposed method first finds the “intersection” of the arm and up torso on the curve skeleton based on the skeleton segmentation. Then we locate another endpoint of the arm skeleton and compute the median point of the arm skeleton. Whether the arm is higher or lower than shoulder can be determined by comparing the “intersection” and median point.
Video based human crowd analysis using machine learning: a survey
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2022
Deevesh Chaudhary, Sunil Kumar, Vijaypal Singh Dhaka
ETH (Luo et al. 2014a) is a pedestrian detection dataset. There are 1,804 images in three video clips in the testing set. The data was taken with a stereo rig mounted on an automobile at a resolution of 640 × 480 pixels and a framerate of 13–14 frames per second. Two scenes ETH and HOTEL are found in ETH. INRIA (Zhang et al. 2016a) is one of the most prominent pedestrian datasets for handmade features-based algorithms. It was gathered as part of a study on detecting upright persons in photos and videos. There are 614 person detections for training and 288 for testing in this set. The majority of the human body in the image is standing, and the height exceeds up to 100 pixels. The Caltech Pedestrian Dataset (Du et al. 2017) is made up of roughly 10 hours of 640 × 48,030 Hz footage captured from a car travelling through ordinary traffic in an urban setting. Around 250,000 frames (in 137 roughly minute-long parts) with a total of 350,000 bounding boxes and 2300 individual pedestrians were annotated. The annotation provides comprehensive occlusion labels and temporal correspondence between bounding boxes. KITTI (Hu et al. 2018a) is a difficult computer vision benchmark that includes a variety of vision tasks. There are 7,481 training photos and 7,518 test images for object detection (vehicle detection, pedestrian detection, and bike detection). The photos have a resolution of 1,240 x 376 pixels.