In this paper, an integrated algorithm to detect humans in thermal imagery was introduced. In recent years, histogram of oriented gradient (HOG) is a quite popular algorithm for person detection in visible imagery. We implement the pedestrian detection in infrared imagery with this algorithm by adjusting the parameters. Simultaneously, we have increased some other geometric characteristics, such as mean contrast, which is used as features for the detection. After analyzing the property of the infrared imagery, which is designed to meet the shortfall of the HOG in infrared imagery, the combined vectors are fed to a linear SVM for object/non-object classification and we get the detector at the same time. After that, the detection window is scanned across the image at multiple positions and scales, which is followed by the combination of the overlapping detections. At last, a pedestrian is described by a final detection, and we have detected the pedestrians in the thermal imagery. Experimental results with OSU Thermal Pedestrian Database are reported to demonstrate the excellent performance of our algorithms.
This paper designs a new control system of high-subdivision for a stepper motor. through which integrated controller and driver of stepper motor on a FPGA chip. The problem of high-subdivision stepping angle is solved by using the technique of FPGA hardware at controllers and drivers of stepper. The number of subdivision can reach 4096, and the step pitch angle be automatically adjusted. This system has also established a PWM generator, a PI controller and a microprocessor IP core, that may further be used of making control chip of stepper motors. Product with the same function, at present, which has this great performance has not been seen in any report. Experimental results indicate that the technique could improve the resolution of a stepper motor and make it run more smoothly.
The vulnerability of machine learning models to Membership Inference Attacks (MIAs) has garnered considerable attention in recent years. These attacks determine whether a data sample belongs to the model's training set or not. Recent research has focused on reference-based attacks, which leverage difficulty calibration with independently trained reference models. While empirical studies have demonstrated its effectiveness, there is a notable gap in our understanding of the circumstances under which it succeeds or fails. In this paper, we take a further step towards a deeper understanding of the role of difficulty calibration. Our observations reveal inherent limitations in calibration methods, leading to the misclassification of non-members and suboptimal performance, particularly on high-loss samples. We further identify that these errors stem from an imperfect sampling of the potential distribution and a strong dependence of membership scores on the model parameters. By shedding light on these issues, we propose RAPID: a query-efficient and computation-efficient MIA that directly \textbf{R}e-lever\textbf{A}ges the original membershi\textbf{P} scores to m\textbf{I}tigate the errors in \textbf{D}ifficulty calibration. Our experimental results, spanning 9 datasets and 5 model architectures, demonstrate that RAPID outperforms previous state-of-the-art attacks (e.g., LiRA and Canary offline) across different metrics while remaining computationally efficient. Our observations and analysis challenge the current de facto paradigm of difficulty calibration in high-precision inference, encouraging greater attention to the persistent risks posed by MIAs in more practical scenarios.
In order to improve degradation efficiency and reduce electrode cost, carbon microspheres (CM) particle electrode was prepared from easily available duckweed as raw materials for electrocatalytic hydrodechlorination. The surface structures and chemical characteristics of CM were regulated by adjusting the pyrolysis temperature. When carbonization temperature was 650 ◦C, the carbon microspheres with high surface area, rich in heteroatom (N, P) and high electrical conductivity are obtained. After Ru catalyst was supported, the Ru/CM-650 as particle electrode form a three-dimensional (3D) electrochemical reaction system exhibits high electrocatalytic performance for DCF hydrodechlorination with dechlorination efficiency of 90% in 150 min, with Ru load was only 1.52wt%. Furthermore, Ru/CM-650 has good stability after repeated use, the reason may be that the synergistic interaction between N and P elements improves the stability of the catalyst.
Virtual Trusted Platform Modules (vTPMs) are widely used in commercial cloud platforms (e.g., VMware Cloud, Google Cloud, and Microsoft Azure) to provide virtual root-of-trust and security services for virtual machines. Unfortunately, current state-of-the-art vTPM implementations for cloud computing cannot provide strong protection for vTPMs at run-time and suffer from poor performance under binding vTPMs to a physical TPM. In this paper, we propose SvTPM, an SGX-based virtual trusted platform module, which provides complete life cycle protection of vTPMs in the cloud and does not rely on the physical TPM. SvTPM provides strong isolation protection so malicious cloud tenants or even cloud administrators cannot access vTPM's private keys or any other sensitive data. In this paper, we implement a prototype of SvTPM, which identifies and solves a couple of critical security challenges for vTPM protection with SGX, such as NVRAM rollback attacks, NVRAM binding attacks, and vTPM rollback attacks. SvTPM also shows how to establish trust between vTPM and SGX Platform. Our performance evaluation shows that the NVRAM launch time of SvTPM is $1700\times$ faster than vTPM built upon hardware TPM. In TPM standard command evaluation, we find that SvTPM incurs negligible performance overhead while providing strong isolation and protection. To our knowledge, SvTPM is the first practical work to solve the critical security challenges of securing vTPM using SGX.
The vulnerability of machine learning models to Membership Inference Attacks (MIAs) has garnered considerable attention in recent years. These attacks determine whether a data sample belongs to the model's training set or not. Recent research has focused on reference-based attacks, which leverage difficulty calibration with independently trained reference models. While empirical studies have demonstrated its effectiveness, there is a notable gap in our understanding of the circumstances under which it succeeds or fails. In this paper, we take a further step towards a deeper understanding of the role of difficulty calibration. Our observations reveal inherent limitations in calibration methods, leading to the misclassification of non-members and suboptimal performance, particularly on high-loss samples. We further identify that these errors stem from an imperfect sampling of the potential distribution and a strong dependence of membership scores on the model parameters. By shedding light on these issues, we propose RAPID: a query-efficient and computation-efficient MIA that directly Re-leverAges the original membershiP scores to mItigate the errors in Difficulty calibration. Our experimental results, spanning 9 datasets and 5 model architectures, demonstrate that RAPID outperforms previous state-of-the-art attacks (e.g., LiRA and Canary offline) across different metrics while remaining computationally efficient. Our observations and analysis challenge the current de facto paradigm of difficulty calibration in high-precision inference, encouraging greater attention to the persistent risks posed by MIAs in more practical scenarios.
Interpersonal trust relationship is an important dimension of interpersonal relationships. With new introduced plots of literature, we can evaluate the environment of characters and predict plot development to some extent. This paper proposes a representation of interpersonal trust relationship based on the fuzzy set theory in the restricted domain of A Dream of Red Mansions. Interpersonal trust degrees are obtained by the comprehensive evaluation based on the fuzzy analytic hierarchy process. This paper proposes a thought that solves the divergences of domain experts, and proposes an approach that establishes the initial trust degree between characters. Based on above, this paper analyzes the trust bias and overall trust degree between characters and mines the relationship between characters and content of the whole trust network of characters. The experiments show that the model is efficient in reflecting the interpersonal trust relationship of A Dream of Red Mansions. The interpersonal trust relationship of A Dream of Red Mansions is modeled and analyzed by the fuzzy set theory, which provides a novel thought on mathematical study of interpersonal trust relationship in literary works.
'/%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)