Existing face anti-spoofing models using deep learning for multi-modality data suffer from low generalization in the case of using variety of presentation attacks such as 2D printing and high-precision 3D face masks. One of the main reasons is that the non-linearity of multi-spectral information used to preserve the intrinsic attributes between a real and a fake face are not well extracted. To address this issue, we propose a multi-modility data based two-stage cascade framework for face anti-spoofing. The proposed framework has two advantages. Firstly, we design a two-stage cascade architecture that can selectively fuse lowlevel and high-level features from different modalities to improve feature representation. Secondly, we use multi-modality data to construct a distance-free spectral on RGB and infrared (IR) to augment the non-linearity of data. The presented data fusion strategy is different from popular fusion approaches, since it can strengthen discrimination ability of network models on physical attribute features than identity structure features under certain constraints. In addition, a multi-scale patch based weighted finetuning strategy is designed to learn each specific local face region. Experimental results show that the proposed framework achieves better performance than other state-of-the-art methods on both benchmark datasets and self-established datasets, especially on multi-material masks spoofing.

An active engine mount (AEM) is an effective technology to improve a vehicle’s noise, vibration, and harshness performance. This paper mainly focuses on the combination experiment and finite element analysis (FEA) for parameter identification of AEMs. Notably, a novel test rig is designed to identify all specific parameters involved in the AEM. Firstly, the static and dynamic stiffness of the main rubber spring are calculated based on structure FEA method. The equivalent piston area and upper chamber volumetric stiffness are also estimated through fluid–structure interaction analysis. Inertia track parameters, involving inertia and linear and nonlinear resistance of the fluid, are identified by a simplified fluid model. These common hydraulic engine mount parameters are all experimentally validated through the original test rig. Besides, the particular components of the electromagnetic AEM, namely actuator parameters, are further estimated by experimental identification utilizing the experimental apparatus. The novel test bench, which exhibited high accuracy, good tightness, and strong versatility, not only simplifies the structure and process of identification plant for passive engine mount parameters, but accommodates the particular AEM ones. The combination method assimilates both the efficiency of FEA and the accuracy of experiment, facilitating the structure design and renovation of AEMs.

Active control engine mounts notably contribute to ensuring superior effectiveness in vibration suppression. The filtered-x-least-mean-squares algorithm, as a benchmark, is widely implemented for cancelation of disturbing engine vibrations. Such an algorithm requires an accurate secondary path estimate to ensure better performance. This study illustrates that incorporating body input point inertance to the active control engine mount model is necessary when accelerometers are utilized in most practical applications. Secondary path estimation errors caused by neglecting body input point inertance are pointed out via the secondary path modeling mismatch theory. Furthermore, the active control engine mount control system is evolved to fit the acceleration transducer applications. On the basis of the improved active control engine mount control system, a novel extended filtered-x-least-mean-squares algorithm based on the acceleration error signal is proposed to adapt to the extended control system. In the end, severe control collapse of secondary path estimation errors caused by neglecting body input point inertance is verified through simulation. Simulated results are presented to validate the performance of the extended filtered-x-least-mean-squares algorithm based on the acceleration error signal. The study demonstrates that the algorithm produces results showing effective vibration isolation. KeywordsActive control engine mount, body input point inertance, secondary path modeling mismatch, extended filtered-x-leastmean-squares algorithm, acceleration error signal Date

No abstract

No abstract