Proposed Detection Method Research Articles

DFGNN: An interpretable and generalized graph neural network for deepfakes detection

Deepfakes are generated using sophisticated deep-learning models to create fake images or videos. As the techniques for creating deepfakes improve, issues like defamation, impersonation, fraud, and misinformation on social media are becoming more prevalent. Existing deep learning-based deepfakes detection models are not interpretable and don’t generalize well when tested across diverse deepfakes generating techniques and datasets. Therefore, the creation of reliable and effective deepfakes detection algorithms is required which are not only generalizable but also interpretable. This paper introduces a novel graph neural network-based architecture to identify hyper-realist deepfake content. Currently, very limited efforts have been done to address the problem of deepfakes detection using graph neural networks. The proposed model is based on the pyramid structure that takes advantage of multi-scale images property by extracting features with progressively smaller spatial sizes as layer depth increases. The method first sliced the image into patches, which are referred to as nodes, and then constructed a graph by connecting the nearest neighbors. To transform and exchange information between all nodes, the proposed model has two basic modules: GraphNet, which uses graph convolution layers to aggregate and update graph information, and FFN, which has linear layers for the transformation of node features. The effectiveness of the method is assessed using the diverse Deepfake Detection Challenge dataset (DFDC), FaceForensics++ (FF++), World Leaders dataset (WLRD), and the Celeb-DF. To demonstrate the generalizability of the proposed method for accurate deepfakes detection, open/close set, cross-set, and cross-corpora evaluations were also performed. The AUC values of 0.98 on FF++, 0.95 on Celeb-DF, 0.92 on DFDC, and 1.00 on most of the sets of WLRD datasets demonstrate the efficacy of the method for identifying manipulated facial images produced using various deepfakes techniques.

Analysis of ultrasonic nonlinearity for bolt hole crack detection in rail joints by wavelet packet decomposition

This paper proposes a crack detection method for rail joint screw holes based on the ultrasonic nonlinearity. Experimental and numerical studies are conducted to validate the performance of the proposed approach. The ultrasonic signals excited and received at the rail head are processed by spectral analysis and Wavelet Packet Decomposition (WPD). The third harmonic can be clearly observed from the spectrum analysis, which is supposed to be the nonlinearity of structures. Further, WPD is used to decompose the frequency components of the received signals. A damage index (DI) is defined based on the energy percentage of the energy in the WPD node containing the third harmonic with respect to the total energy of the received signal. It is found that the DI relating to the third harmonic increases with an approximately linear relationship when the crack grows bigger. It can provide a basis for the crack detection and quantification for the rail joint screw holes. A finite element model of the experimentally tested rail is also developed and numerical studies are conducted to verify the ultrasonic nonlinearity in the rail structure. Numerical results have the same trend with the experimental results, which verifies the effectiveness of the proposed method for crack detection in rail joints based on the high order ultrasonic nonlinear harmonics.

FENet: feature enhancement network for land cover classification

ABSTRACT Extracting mask information of buildings and water areas from high resolution remote sensing images is beneficial to monitoring and management of urban development. However, due to different times, different geographical locations and different remote sensing acquisition angles, water areas and buildings will feed back different spectral information. Existing semantic segmentation methods do not pay enough attention to channel information, and the feature information extracted by downsampling is relatively abstract, which is easy to cause the loss of some details in high-resolution images under complex scenes, leading to the misjudgement of buildings and waters. To solve the existing problems, feature enhancement network (FENet) for high-resolution remote sensing image segmentation of buildings and water areas is proposed. By paying more attention to the characteristic information of the passage, the probability of misjudgement of buildings and waters can be reduced and their edge contour information can be enhanced. The self-attention feature module proposed in this paper encodes the context information and transmits it to the local features, and establishes the relationship between channels through the channel feature enhancement module to reduce the loss of channel feature information. The feature fusion module fuses feature information of different scales in space and outputs more detailed prediction images. Comparative experiments show that this model is superior to the existing classical semantic segmentation model. Compared with the existing models, the proposed method can achieve 2% improvement than PSPNet on the indicator MIoU, and the final MIoU reaches 82.85% for land cover dataset. This study demonstrates the advantages of our proposed method in land cover classification and detection.

Wavelet image scattering based glaucoma detection

BackgroundThe ever-growing need for cheap, simple, fast, and accurate healthcare solutions spurred a lot of research activities which are aimed at the reliable deployment of artificial intelligence in the medical fields. However, this has proved to be a daunting task especially when looking to make automated diagnoses using biomedical image data. Biomedical image data have complex patterns which human experts find very hard to comprehend. Against this backdrop, we applied a representation or feature learning algorithm: Invariant Scattering Convolution Network or Wavelet scattering Network to retinal fundus images and studied the the efficacy of the automatically extracted features therefrom for glaucoma diagnosis/detection. The influence of wavelet scattering network parameter settings as well as 2-D channel image type on the detection correctness is also examined. Our work is a distinct departure from the usual method where wavelet transform is applied to pre-processed retinal fundus images and handcrafted features are extracted from the decomposition results. Here, the RIM-ONE DL image dataset was fed into a wavelet scattering network developed in the Matlab environment to achieve a stage-wise decomposition process called wavelet scattering of the retinal fundus images thereby, automatically learning features from the images. These features were then used to build simple and computationally cheap classification algorithms.ResultsMaximum detection correctness of 98% was achieved on the held-out test set. Detection correctness is highly sensitive to scattering network parameter setting and 2-D channel image type.ConclusionA superficial comparison of the classification results obtained from our work and those obtained using a convolutional neural network underscores the potentiality of the proposed method for glaucoma detection.

Wrap-Around Based Frequency-Varying Viterbi Detection for Windowed OFDM Systems in Doubly Selective Fading Channels

In doubly selective fading channels, the conventional orthogonal frequency division multiplexing (OFDM) system suffers high inter-carrier interferences (ICIs). To deal with the ICIs, combining the time-domain windowing and Viterbi detection, we propose a wrap-around based frequency-varying Viterbi detection scheme for windowed OFDM systems. In a doubly selective fading channel, we theoretically derive the input-output relationship of windowed OFDM systems. It is observed that by selecting a proper window function, the equivalent channel matrix is sparse. Therefore, the severe ICIs can be significantly mitigated by our proposed wrap-around based frequency-varying Viterbi detection scheme. It is shown by simulation results that our proposed scheme achieves the performance very close to that of the maximum likelihood detector, even when the normalized Doppler frequency is as high as 0.667. Furthermore, the computational complexity of our proposed detection method is linear to the number of subcarriers, given the maximum iteration times and the constraint length.

A New Electrical Resistance Measurement Method of Carbon Fiber-Reinforced Polymer Tubes Based on C⁴D Technique

This work introduces the capacitively coupled contactless conductivity detection ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\text{C}^{{4}}\text{D}$ </tex-math></inline-formula> ) technique into nondestructive testing (NDT) of carbon fiber-reinforced polymer (CFRP) and proposes a new resistance measurement method of CFRP tubes. The proposed method does not require direct contact between the electrodes and the carbon fiber. Therefore, the electrical resistance of CFRP can be measured without the treatments required by traditional electrical methods that may lead to surface or structural damage. The equivalent circuit model of the new method is established, and the challenges are analyzed. Simulation and experiment are conducted with woven fabric CFRP tubes to verify the feasibility of the proposed method, including the range of workable frequency, the effectiveness of the circuit model, and the capability of damage detection. Research results show that a relatively high working frequency is required to accurately measure the interested resistance and sense the resistance change. By selecting a workable frequency, the effectiveness of the circuit model is verified. Research results indicate that the stray capacitance between the electrodes has little effect on resistance measurement, and the capacitor–resistor–capacitor model is still applicable. Research results also verify the feasibility of the proposed method for damage detection. It is able to sense the resistance change resulted from the damage of CFRP, either the conductivity changes in simulation or actual cutting damage in experiment.

Reference-free fatigue crack detection using deep long short-term memory network (DLSTM) and nonlinear ultrasonic modulation

Nonlinear ultrasonic modulation is sensitive to fatigue crack, but a reference signal or user-specified threshold is often required for crack diagnosis, easily causing false alarms in noisy environments. In this study, a reference-free damage detection method was developed by applying a deep long short-term memory network (DLSTM) to nonlinear ultrasonic modulation signals. First, an ultrasonic signal was generated and measured using piezoceramic ultrasonic transducers. Subsequently, a DLSTM network was constructed and trained to learn the inherent sequential patterns of the measured ultrasonic signals. Then, an absolute damage index (ADI) was defined and computed using only the current ultrasonic signal without any reference ultrasonic signal obtained from the intact condition. Finally, the crack was automatically detected using the ADI and without any user-specified threshold. The performance of the proposed method was examined using data from a submerged floating tunnel model and an actual long-span bridge. The results highlight the feasibility of the proposed method for automatic fatigue crack detection.

Adaptive Reflection Detection and Control Strategy of Pointer Meters Based on YOLOv5s

Reflective phenomena often occur in the detecting process of pointer meters by inspection robots in complex environments, which can cause the failure of pointer meter readings. In this paper, an improved k-means clustering method for adaptive detection of pointer meter reflective areas and a robot pose control strategy to remove reflective areas are proposed based on deep learning. It mainly includes three steps: (1) YOLOv5s (You Only Look Once v5-small) deep learning network is used for real-time detection of pointer meters. The detected reflective pointer meters are preprocessed by using a perspective transformation. Then, the detection results and deep learning algorithm are combined with the perspective transformation. (2) Based on YUV (luminance-bandwidth-chrominance) color spatial information of collected pointer meter images, the fitting curve of the brightness component histogram and its peak and valley information is obtained. Then, the k-means algorithm is improved based on this information to adaptively determine its optimal clustering number and its initial clustering center. In addition, the reflection detection of pointer meter images is carried out based on the improved k-means clustering algorithm. (3) The robot pose control strategy, including its moving direction and distance, can be determined to eliminate the reflective areas. Finally, an inspection robot detection platform is built for experimental study on the performance of the proposed detection method. Experimental results show that the proposed method not only has good detection accuracy that achieves 0.809 but also has the shortest detection time, which is only 0.6392 s compared with other methods available in the literature. The main contribution of this paper is to provide a theoretical and technical reference to avoid circumferential reflection for inspection robots. It can adaptively and accurately detect reflective areas of pointer meters and can quickly remove them by controlling the movement of inspection robots. The proposed detection method has the potential application to realize real-time reflection detection and recognition of pointer meters for inspection robots in complex environments.

An unsupervised learning damage diagnosis method based on virtual impulse response function and time series models

As it contains the inherent properties of the structure and is insensitive to variation of the input, the transmissibility function (TF) is widely used for structural damage detection under ambient excitation. However, in TF-based methods, the selection of the optimal damage sensitivity band often requires access to vibration data under different operating conditions, which is not conducive to unsupervised learning. In order to accommodate unsupervised learning and extract more damage information, this paper proposes a data-driven unsupervised damage detection method based on the virtual impulse response function (VIRF) and time series model (TSM). In this method, the VIRF is used as a damage-sensitive feature (DSF), the TSM is used for low-dimensional feature extraction of the VIRF, and the distance metric of the model coefficients is used to detect the occurrence of anomalies. The study analyses the type of TSM applied to VIRF and uses the parameter adjusted cosine similarity (PACS) as a more accurate statistical distance tool for the model coefficients. The application of two well-known benchmark structure experiments verifies the accuracy and practicality of the proposed method in damage detection under minor operational variations and ambient excitations. The comparative analysis and detection results indicate that the proposed method is advantageous in qualitatively quantifying the extent of the damage. Structural anomalies can be detected in both linear and nonlinear damage scenarios.

STCD-EffV2T Unet: Semi Transfer Learning EfficientNetV2 T-Unet Network for Urban/Land Cover Change Detection Using Sentinel-2 Satellite Images

Change detection in urban areas can be helpful for urban resource management and smart city planning. The effects of human activities on the environment and ground have gained momentum over the past decades, causing remote sensing data sources analysis (such as satellite images) to become an option for swift change detection in the environment and urban areas. We proposed a semi-transfer learning method of EfficientNetV2 T-Unet (EffV2 T-Unet) that combines the effectiveness of composite scaled EfficientNetV2 T as the first path or encoder for feature extraction and convolutional layers of Unet as the second path or decoder for reconstructing the binary change map. In the encoder path, we use EfficientNetV2 T, which was trained by the ImageNet dataset. In this research, we employ two datasets to evaluate the performance of our proposed method for binary change detection. The first dataset is Sentinel-2 satellite images which were captured in 2017 and 2021 in urban areas of northern Iran. The second one is the Onera Satellite Change Detection dataset (OSCD). The performance of the proposed method is compared with YoloX-Unet families, ResNest-Unet families, and other well-known methods. The results demonstrated our proposed method’s effectiveness compared to other methods. The final change map reached an overall accuracy of 97.66%.

Android malware detection based on sensitive patterns

In recent years, the rapid increase in the number and type of Android malware has brought great challenges and pressure to malware detection systems. As a widely used method in android malware detection, static detecting has been a hot topic in academia and industry. However, in order to improve the accuracy of detection, the existing static detecting methods sacrifice the excessively high analysis complexity and time cost. Moreover, the correlation between static features leads to redundancy of a large amount of data. Therefore, this paper proposes a static detecting method of Android malware based on sensitive pattern. It uses an improved FP-growth algorithm to mine frequent combinations of sensitive permissions and API calls in malicious apps and benign apps, which avoids the generation of redundant information. In addition, this paper adopts multi-layered gradient boosting decision trees algorithm to train the detection model. And a dual similarity combination method is proposed to measure the similarity between different sensitive patterns. The experimental results show that our proposed detection method has high accuracy and great generalization ability.

Research of energy characteristics and fault position detection for cracked rotor system

Early crack faults will introduce subtle local changes in the whole rotor system, which may bring difficulties to the fault detection process due to external disturbing factors. Meanwhile, judging the damaged position of rotor systems might be tough only through displacement or velocity. To overcome these issues, the harmonic balance-energy power flow analysis (HB-EPFA) is presented to investigate the rotor system with a transverse crack. Besides, a novel indicator (Time-averaged power flow) is proposed to evaluate the crack faults. For the energy characteristic analysis, the cracked rotor appears a special amplitude-jumping phenomenon, even in the case of shallow crack depth. The influence of crack depth on power flow values under different rotating speeds is obtained. Furthermore, the existence of cracks will lead to the rotation shaft presenting an abrupt energy-changing zone, which can be used to detect crack position. Relevant experiments are applied to validate the proposed detection method, and the experimental results agree with the theoretical analysis. The conclusions achieved in this paper can provide a reference for fault diagnosis of rotor systems in engineering practice.

A Kitchen Standard Dress Detection Method Based on the YOLOv5s Embedded Model

In order to quickly and accurately detect whether a chef is wearing a hat and mask, a kitchen standard dress detection method based on the YOLOv5s embedded model is proposed. Firstly, a complete kitchen scene dataset was constructed, and the introduction of images for the wearing of masks and hats allows for the low reliability problem caused by a single detection object to be effectively avoided. Secondly, the embedded detection system based on Jetson Xavier NX was introduced into kitchen standard dress detection for the first time, which accurately realizes real-time detection and early warning of non-standard dress. Among them, the combination of YOLOv5 and DeepStream SDK effectively improved the accuracy and effectiveness of standard dress detection in the complex kitchen background. Multiple sets of experiments show that the detection system based on YOLOv5s has the highest average accuracy of 0.857 and the fastest speed of 31.42 FPS. Therefore, the proposed detection method provided strong technical support for kitchen hygiene and food safety.

Detect Orientation of Symmetric Objects from Monocular Camera to Enhance Landmark Estimations in Object SLAM

Object simultaneous localization and mapping (SLAM) introduces object-level landmarks to the map and helps robots to further perceive their surroundings. As one of the most preferred landmark representations, ellipsoid has a dense mathematical expression and can represent the occupied space of objects with high accuracy. However, the orientations of ellipsoid approximations often fail to coincide with the orientation of objects. To further improve the performance of object SLAM systems with ellipsoid landmarks, we innovatively propose a strategy that first extracts the orientations of those symmetric human-made objects in a single frame and then implements the results of the orientation as a back-end constraint factor of the ellipsoid landmarks. Experimental results obtained show that, compared with the baseline, the proposed orientation detection method can reduce the orientation error by more than 46.5% in most tested datasets and improves the accuracy of mapping. The average translation, rotation and shape error improved by 63.4%, 61.7% and 42.4%, respectively, compared with quadric-SLAM. With only 9 ms additional time cost of each frame, the object SLAM system integrated with our proposed method can still run in real time.

Peanut kernel integrity detection based on deep learning convolution neural network

The edge features of peanut grain image are not considered in peanut grain integrity detection, and there are many noises, resulting in low accuracy of peanut grain integrity detection and poor effect of image edge noise removal. Therefore, this paper designs a peanut seed integrity detection method based on deep learning convolution neural network. The edge contour of peanut grain image is closed, the edge contour curve characteristics of peanut grain image are extracted, the noise area of peanut grain edge image is determined according to the filter window of Gaussian kernel function, and the wavelet descriptor in contour description operator is used to reduce the edge noise of peanut grain image. Input the image into the convolution neural network, update the weight by gradient descent method, construct the peanut seed integrity detection model, output the peanut seed integrity detection results, and realize the peanut seed integrity detection. The experimental results show that the proposed detection method can improve the accuracy of peanut grain detection and has certain feasibility.

Sea-Surface Small Target Detection Based on Four Features Extracted by FAST Algorithm

On account of current algorithm and parameter design difficulties and low detection accuracy in feature extractions of small target detections in sea clutter environment, this paper proposes a correspondingly improved four feature extraction method by FAST. After the short-time Fourier transform is applied, a time–frequency distribution spectrogram of original data is generated. Candidate feature points (CFP) are first extracted by FAST algorithm, and then a four-feature extraction is implemented with FAST and DBSCAN combined. The feature distinction is enhanced through a feature optimization. Upon the construction of the four-dimensional feature vectors, XGBoost classifier algorithm classifies and detects these feature vectors. The genetic algorithm optimizes the hyperparameters in XGBoost and updates the decision threshold in real time to control the detection method’s false alarm rate. The IPIX dataset is employed for experimental verification. Verification results confirm that this proposed detection method has better performance than several other currently used detection methods. The detection performance is improved by 7% and 13.8% when observation time is set at 0.512 s and 1.024 s, respectively.

Application of LSTM Method Combined with Feature Optimization in Chiller Failure Detection

Aiming at the characteristics of a long short-term memory network (LSTM) which is suitable for processing high-dimensional, strongly coupled, and highly time-dependent data, it combines the advantages of feature selection to reduce the difficulty of learning tasks and improve the performance of model fault diagnosis. This paper proposes an LSTM method combining sequential floating forward search with integrated feature selection for chiller sensor deviation fault detection. The detection results of the proposed method are compared with those of the single feature selection method, and it is concluded that the detection efficiency of the proposed method in chiller sensor deviation fault detection is significantly better than that of the single feature selection method.

Second-order transient-extracting S transform for fault feature extraction in rolling bearings

• STEST is proposed to rectify the GD bias in time extracting S-transform. • Time-frequency representation of fault periods can be concentrated by STEST. • Two experiments are performed to detect faults in bearings. Intelligent fault diagnosis methods can obtain promising results in ensuring the safety and reliability of key parts of rotating machinery. However, the problems are the insufficient amount of data during equipment acceptance period and the assumption that the collected data are high quality which directly affects the reliability of promising results. To solve the above problems, based on the characteristics of fault features, a time-frequency-based method is introduced to analyze the impulse components. Nevertheless, the performance of the time-frequency method is deeply relies on the selection of the window length. To avoid the influence of uncertain parameters, an accurate time-frequency analysis method named the second-order transient-extracting S transform based on the S-transform is proposed in this paper. The proposed method not only rectifies the group delay bias but also produces a highly concentrated time-frequency representation even in noise-surrounded and irrelevant components. The effectiveness of the proposed method for monitoring the health of key parts health is verified through simulated and experimental investigations. The accuracy of the proposed method in feature detection is higher than that of other methods.

Fast, specific, and ultrasensitive antibiotic residue detection by monolayer WS2-based field-effect transistor sensor.

Antibiotic residues cause increasing concern in environmental ecology and public health, which needs efficient analysis strategy for monitoring and control. In this study, a fast, specific, and ultrasensitive sensor based on field-effect transistor (FET) has been proposed for the detection of ampicillin (AMP). The sensor involves monolayer tungsten disulfide (WS2) nanosheet as the sensing channel, single-stranded DNA (ssDNA) as the sensing probe, and gold nanoparticle (Au NP) as the linker. The WS2/Au/ssDNA FET sensor responds rapidly to AMP in a wide linear detection range (10-12-10-6 M) and has low limit of detection (0.556pM), which meets the permissible standards of AMP in water and food. The sensing mechanism study suggests that the excellent sensor response results from the increased number of negative charges in the Debye length and the consequent accumulation of holes in WS2 channel after the addition of AMP. Moreover, satisfactory sensing performance was confirmed in real water samples, indicating the potential application of the proposed method in practical AMP detection. The reported FET sensing strategy provides new insights in antibiotic analysis for risk assessment and control.

G-Rep: Gaussian Representation for Arbitrary-Oriented Object Detection

Typical representations for arbitrary-oriented object detection tasks include the oriented bounding box (OBB), the quadrilateral bounding box (QBB), and the point set (PointSet). Each representation encounters problems that correspond to its characteristics, such as boundary discontinuity, square-like problems, representation ambiguity, and isolated points, which lead to inaccurate detection. Although many effective strategies have been proposed for various representations, there is still no unified solution. Current detection methods based on Gaussian modeling have demonstrated the possibility of resolving this dilemma; however, they remain limited to OBB. To go further, in this paper, we propose a unified Gaussian representation called G-Rep to construct Gaussian distributions for OBB, QBB, and PointSet, which achieves a unified solution to various representations and problems. Specifically, PointSet- or QBB-based object representations are converted into Gaussian distributions and their parameters are optimized using the maximum likelihood estimation algorithm. Then, three optional Gaussian metrics are explored to optimize the regression loss of the detector because of their excellent parameter optimization mechanisms. Furthermore, we also use Gaussian metrics for sampling to align label assignment and regression loss. Experimental results obtained on several publicly available datasets, such as DOTA, HRSC2016, UCAS-AOD, and ICDAR2015, show the excellent performance of the proposed method for arbitrary-oriented object detection.