Control Flow Integrity (CFI) provides a strong protection against modern control-flow hijacking attacks. However, performance and compatibility issues limit its adoption. We propose new practical realistic method called CCFIR (Compact Randomization), which addresses the main barriers to CFI collects all legal targets of indirect control-transfer instructions, puts them into dedicated "Springboard section" in random order, then limits transfers flow only them. Using Springboard section for...

Fuzz testing has proven successful in finding security vulnerabilities large programs. However, traditional fuzz tools have a well-known common drawback: they are ineffective if most generated malformed inputs rejected the early stage of program running, especially when target programs employ checksum mechanisms to verify integrity inputs. In this paper, we present TaintScope, an automatic fuzzing system using dynamic taint analysis and symbolic execution techniques, tackle above problem....

Laboratory micro-CT systems, although limited by beam hardening effect and instability of the source, have been utilized to measure mineral density in combination with specific image processing methods. However, few attempts made accurately determine profiles dentine due lack suitable calibration standards. The aim this study was develop a method evaluate including changes associated dentinal caries. A series K(2)HPO(4) solution phantoms concentration range between 0 0.9 g cm(-3)--coupled...

Heap spraying is an attack technique commonly used in hijacking browsers to download and execute malicious code. In this attack, attackers first fill a large portion of the victim process's heap with Then they exploit vulnerability redirect control attackers' code on heap. Because location injected not exactly predictable, traditional heap-spraying attacks need inject huge amount executable increase chance success. Injected usually includes lots NOP-like instructions leading shellcode....

Recently, a variety of acoustic tasks and related applications arised. For many tasks, the labeled data size may be limited. To handle this problem, we propose an unsupervised pre-training method using Transformer based encoder to learn general robust high-level representation for all tasks. Experiments have been conducted on three kinds tasks: speech emotion recognition, sound event detection translation. All experiments shown that its own training can significantly improve performance....

A high-accuracy calibration technique using a white paper and front coating plane mirror is proposed in this for line-structured light vision sensors. This method shows advantages two aspects. First, can gain very high-quality stripe due to its approximate ideal diffuse Lambertian sheet, which overcomes the problems associated with strong reflecting serious burrs of on conventional rigid targets. Second, based lithographic feature points, we obtain bilateral symmetric structure similar...

The depth measurement error of bionic eyes consisting two active cameras is derived by considering the eyes’ rotation angle errors and image feature extraction errors. effect factors are obtained from calculation formula error. Based on analysis these factors, some effective guidelines for proposed to reduce suggest keeping long baseline, observing target as nearly possible, controlling with same angular velocity Z w axis if possible. simulation experiments practical in platform validate...

An extrinsic parameters calibration method of multi-cameras with non-overlapping fields view (FOV) using laser scanning is presented. Firstly, two lasers are mounted on a multi-degree-of-freedom manipulator and can scan objects freely by the projected line-structured light. Then, controlling movement manipulator, light into field one multi-cameras, plane equation in camera coordinate frame calibrated target. The moved several times small amplitude to change position structured vision...

This paper presents the details of our system designed for Task 1 Multimodal Information Based Speech Processing (MISP) Challenge 2021. The purpose is to leverage both audio and video information improve environmental robustness far-field wake word spotting. In proposed system, firstly, we take advantage speech enhancement algorithms such as beamforming weighted prediction error (WPE) address multi-microphone conversational audio. Secondly, several data augmentation techniques are applied...

This paper presents the development and evaluation of ChatHome, a domain-specific language model (DSLM) designed for intricate field home renovation. Considering proven competencies large models (LLMs) like GPT-4 escalating fascination with renovation, this study endeavors to reconcile these aspects by generating dedicated that can yield high-fidelity, precise outputs relevant renovation arena. ChatHome's novelty rests on its methodology, fusing domain-adaptive pretraining instruction-tuning...

Although the quality of reconstructed results can be improved with increment number measurements, scale matrices involved in reconstruction fluorescence molecular tomography (FMT) will become larger, which leads to poor efficiency process tomographic image reconstruction. In this paper, we proposed a new method for FMT based on compressed sensing, scheme grouped sources is incorporated. The forward equations are implemented using finite element (FEM). model formulated under framework sensing...

Recently, a variety of acoustic tasks and related applications arised. For many tasks, the labeled data size may be limited. To handle this problem, we propose an unsupervised pre-training method using Transformer based encoder to learn general robust high-level representation for all tasks. Experiments have been conducted on three kinds tasks: speech emotion recognition, sound event detection translation. All experiments shown that its own training can significantly improve performance....

In many measurement applications using multi-cameras, there is often a non-overlapping field of view (nFOV) between multi-cameras. It benefits greatly from having all the cameras unified in single coordinate frame by means common spatial constraints. However, it very difficult to establish constraints for accurate and quick calibration extrinsic parameters multi-cameras case long working distance, asymmetric angle or limited space. To overcome these issues, we propose flexible method camera...

A simple binocular visual equipment is designed in this paper, which mainly composed by two CCD cameras and four stepping motors. Firstly, the cameras' initial location determined based on rough moving to limiting position, precise driving image center aligning with fitted ellipse center. Then, transformation matrix between coordinate systems of motion module end camera calibrated eye-in-hand calibration method, adopts multi frame images a method Levenberg-Marquardt iterative algorithm...