Dynamic hand gesture acting as a semaphoric gesture is a practical and intuitive mid-air gesture interface. Nowadays benefiting from the development of deep convolutional networks, the gesture recognition has already achieved a high accuracy, however, when performing a dynamic hand gesture such as gestures of direction commands, some unintentional actions are easily misrecognized due to the similarity of the hand poses. This hinders the application of dynamic hand gestures and cannot be solved by just improving the accuracy of the applied algorithm on public datasets, thus it is necessary to study such problems from the perspective of human-computer interaction. In this article, two methods are proposed to avoid misrecognition by introducing activation delay and using asymmetric gesture design. First the temporal process of a dynamic hand gesture is decomposed and redefined, then a realtime dynamic hand gesture recognition system is built through a two-dimensional convolutional neural network. In order to investigate the influence of activation delay and asymmetric gesture design on system performance, a user study is conducted and experimental results show that the two proposed methods can effectively avoid misrecognition. The two methods proposed in this article can provide valuable guidance for researchers when designing realtime recognition system in practical applications.
Mobile crowd sensing (MCS) is a new sensing paradigm that leverages participatory sensing data from mobile devices for accomplishing large-scale sensing tasks. Incentivizing device owners to contribute high-quality sensing data is a prerequisite for the success of MCS services. In this paper, we first propose a pre-contracting incentive mechanism that involves the participation of not only the device owners located in close proximity to Point of Interests (PoIs) but also the device owners that are going to pass through those locations. Furthermore, the quality of sensing data is guaranteed through the use of redundancy. In particular, sensing data from multiple device owners is processed and compared at an edge side (i.e., base station) so as to detect the measurement error at the proximity of data sources. Simulation results confirm that the proposed incentive mechanism is efficient in terms of improving the total utility.
Based on the theory of Human Factor Engineering, Modular design and Interface technology, the paper put forward rapid design plan and prototype system for engine under the guidance of those theories. The innovation of this system was fulfilling transmitting parameters of engine by using interface technology, and achieving the application of module structure by using interface template. The page design of the system was guided by color collocation theory, readability and operability. System evaluation fully embodied the humanization design, and emphasized user's experience.
Convolutional network models (CNN) are very vulnerable to adversarial samples, which poses a serious challenge to the security of CNN models. Based on the task of CNN's modulation and identification of communication signals, we propose a white-box attack algorithm, the shortest distance attack method (SD-Alg), which can generate extremely small disturbances and greatly reduce the classification performance of the model. Experiments show that our algorithm excels in attack success rate, running time and adversarial perturbation size among the same type of algorithms.
Noise identification is one of the most important issues in radar and sonar engineering, in which Kolmogorov–Smirnov (KS) statistic is widely adopted. Although KS statistic is the most reliable method till now, it does not have high precision as people believe. The performance of KS statistic is studied in this work, while an experiment is carried out to show the precision of KS statistic, including both test and recognition. The result could help researchers decreasing error of using KS statistic in the future.
Virtual reality (VR) technology provides highly immersive depth perception experiences; nevertheless, stereoscopic visual fatigue (SVF) has become an important factor currently hindering the development of VR applications. However, there is scant research on the underlying neural mechanism of SVF, especially those induced by VR displays, which need further research. In this paper, a Go/NoGo paradigm based on disparity variations is proposed to induce SVF associated with depth perception, and the underlying neural mechanism of SVF in a VR environment was investigated. The effects of disparity variations as well as SVF on the temporal characteristics of visual evoked potentials (VEPs) were explored. Point-by-point permutation statistical with repeated measures ANOVA results revealed that the amplitudes and latencies of the posterior VEP component P2 were modulated by disparities, and posterior P2 amplitudes were modulated differently by SVF in different depth perception situations. Cortical source localization analysis was performed to explore the original cortex areas related to certain fatigue levels and disparities, and the results showed that posterior P2 generated from the precuneus could represent depth perception in binocular vision, and therefore could be performed to distinguish SVF induced by disparity variations. Our findings could help to extend an understanding of the neural mechanisms underlying depth perception and SVF as well as providing beneficial information for improving the visual experience in VR applications.
Data imbalance is a crucial factor that limits the performance of automatic defect recognition systems in castings. The bias and deterioration of the model are generated by massive normal samples and minor defect samples. Traditional re-sampling methods randomly change the data distribution and ignore the significant intra-class difference among all normal samples. Therefore, this paper proposes a distribution-preserving under-sampling method for imbalance defect-recognition in castings. In detail, our method divides all normal samples into several sub-groups by cluster analysis and reassembles them into some balance datasets, which makes the normal samples in all balance datasets have an identical distribution with the original imbalance dataset. Finally, experiments on our dataset with 3260 images indicate that the proposed method achieves a 0.816 AUC (area under curve) score, which demonstrates significant advantages compared to cost-sensitive learning and re-sampling methods.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)
'%3e%3cpath fill='%23012169' d='M0 0v30h60V0z'/%3e%3cpath stroke='%23fff' stroke-width='6' d='m0 0 60 30m0-30L0 30'/%3e%3cpath stroke='%23C8102E' stroke-width='4' d='m0 0 60 30m0-30L0 30' clip-path='url(%23b)'/%3e%3cpath stroke='%23fff' stroke-width='10' d='M30 0v30M0 15h60'/%3e%3cpath stroke='%23C8102E' stroke-width='6' d='M30 0v30M0 15h60'/%3e%3c/g%3e%3c/svg%3e)
'/%3e%3c/g%3e%3cuse xlink:href='%23d' x='45.714'/%3e%3cuse xlink:href='%23e' transform='matrix(-1 0 0 1 479.792 0)'/%3e%3cg id='f' fill='%23fff'%3e%3cpath fill='%23039' d='M232.636 202.406v.005c0 2.212.85 4.227 2.212 5.69 1.365 1.466 3.245 2.377 5.302 2.377 2.067 0 3.944-.905 5.303-2.365 1.358-1.459 2.202-3.472 2.202-5.693v-10.768l-14.992-.013-.028 10.766'/%3e%3ccircle cx='236.074' cy='195.735' r='1.486'/%3e%3ccircle cx='244.392' cy='195.742' r='1.486'/%3e%3ccircle cx='240.225' cy='199.735' r='1.486'/%3e%3ccircle cx='236.074' cy='203.916' r='1.486'/%3e%3ccircle cx='244.383' cy='203.905' r='1.486'/%3e%3c/g%3e%3cuse xlink:href='%23f' y='-26.016'/%3e%3cuse xlink:href='%23f' x='-20.799'/%3e%3cuse xlink:href='%23f' x='20.745'/%3e%3cuse xlink:href='%23f' y='25.784'/%3e%3c/svg%3e)