Feature Matching Research Articles

A Digital Animation Generation Model Based on Cycle Adversarial Neural Network

This paper proposes a digital animation generation model based on Cycle Adversarial Neural Network (CycleGAN). Compared with the classical CycleGAN, this research presents a multi-attention approach to enhance the network’s generalization. Specifically, a style enhancement module and a style cross-attention mechanism are introduced into the generator network structure, which enables the model to better parse the structural information of the content image and realize the accurate matching of content features and style features. Furthermore, the introduction of a multi-scale discriminator with a fused attention mechanism enhances the preservation of content information from the source image in the output animated image. The experiments conducted in this study demonstrate that the model proposed in this paper exhibits superior performance in generating digital animation. This model not only enhances the realism and variety of the generated effects, but also achieves significant improvements in the coherence and stability of long time-series animation. The contribution of this paper lies in the introduction of the multi-attention mechanism, which enhances the generalization of the network and is of great theoretical and practical significance for the development of the field of digital animation generation.

An Autonomous Navigation Method for Planetary Rover Based on Multi-modal Fusion and Multi-factor Graph Optimization

Planetary surface is a complex unstructured environment with strong light, strong shadows, weak textures and numerous obstacles. There is no GPS signal support and satellite networking support on the planetary surface. The limited computer and sensor capabilities of planetary rover have a significant gap from the ground autonomous driving configuration, which presents a great challenge to the autonomous navigation methods of planetary rover. Extreme lighting conditions will increase the camera mismatching rate, and poor texture will increase the lidar matching error, which will reduce the estimation accuracy. Moreover, camera and lidar produce great motion distortion in the rugged terrain and bumpy environment, which leads to a large increase in the cumulative error of odometer. Camera and lidar are not suitable as the main sensor of SLAM algorithm in unstructured environment because they are sensitive to the change of environment. Therefore, an autonomous navigation method for planetary rover based on multi-modal fusion and multi-factor graph optimization is presented in this paper. The IMU (Inertial Measurement Unit) odometer node is added to the lidar odometer and visual odometer thread, and the IMU is taken as the central node to build a multi-factor graph optimization model. A multi-factor graph optimization model and a strong adaptive constraint strategy with reasonable weights are constructed. The pose estimation of other sensors is used to constrain IMU bias, meanwhile historical rut tracking, obstacle contour matching and skyline feature matching are introduced. Finally, motion prediction is achieved in the IMU odometer node. Simulation results and field tests show that this method can effectively cope with the unstructured environment on planetary surface, which achieves adaptive and robust autonomous navigation of planetary rover.

RIGA: Rotation-Invariant and Globally-Aware Descriptors for Point Cloud Registration.

Successful point cloud registration relies on accurate correspondences established upon powerful descriptors. However, existing neural descriptors either leverage a rotation-variant backbone whose performance declines under large rotations, or encode local geometry that is less distinctive. To address this issue, we introduce RIGA to learn descriptors that are Rotation-Invariant by design and Globally-Aware. From the Point Pair Features (PPFs) of sparse local regions, rotation-invariant local geometry is encoded into geometric descriptors. Global awareness of 3D structures and geometric context is subsequently incorporated, both in a rotation-invariant fashion. More specifically, 3D structures of the whole frame are first represented by our global PPF signatures, from which structural descriptors are learned to help geometric descriptors sense the 3D world beyond local regions. Geometric context from the whole scene is then globally aggregated into descriptors. Finally, the description of sparse regions is interpolated to dense point descriptors, from which correspondences are extracted for registration. To validate our approach, we conduct extensive experiments on both object- and scene-level data. With large rotations, RIGA surpasses the state-of-the-art methods by a margin of 8 ° in terms of the Relative Rotation Error on ModelNet40 and improves the Feature Matching Recall by at least 5 percentage points on 3DLoMatch.

A Mars Local Terrain Matching Method Based on 3D Point Clouds

To address the matching challenge between the High Resolution Imaging Science Experiment (HiRISE) Digital Elevation Model (DEM) and the Mars Orbiter Laser Altimeter (MOLA) DEM, we propose a terrain matching framework based on the combination of point cloud coarse alignment and fine alignment methods. Firstly, we achieved global coarse localization of the HiRISE DEM through nearest neighbor matching of key Intrinsic Shape Signatures (ISS) points in the Fast Point Feature Histograms (FPFH) feature space. We introduced a graph matching strategy to mitigate gross errors in feature matching, employing a numerical method of non-cooperative game theory to solve the extremal optimization problem under Karush–Kuhn–Tucker (KKT) conditions. Secondly, to handle the substantial resolution disparities between the MOLA DEM and HiRISE DEM, we devised a smoothing weighting method tailored to enhance the Voxelized Generalized Iterative Closest Point (VGICP) approach for fine terrain registration. This involves leveraging the Euclidean distance between distributions to effectively weight loss and covariance, thereby reducing the results’ sensitivity to voxel radius selection. Our experiments show that the proposed algorithm improves the accuracy of terrain registration on the proposed Curiosity landing area’s, Mawrth Vallis, data by nearly 20%, with faster convergence and better algorithm robustness.

Improving Target Geolocation Accuracy with Multi-View Aerial Images in Long-Range Oblique Photography

Target geolocation in long-range oblique photography (LOROP) is a challenging study due to the fact that measurement errors become more evident with increasing shooting distance, significantly affecting the calculation results. This paper introduces a novel high-accuracy target geolocation method based on multi-view observations. Unlike the usual target geolocation methods, which heavily depend on the accuracy of GNSS (Global Navigation Satellite System) and INS (Inertial Navigation System), the proposed method overcomes these limitations and demonstrates an enhanced effectiveness by utilizing multiple aerial images captured at different locations without any additional supplementary information. In order to achieve this goal, camera optimization is performed to minimize the errors measured by GNSS and INS sensors. We first use feature matching between the images to acquire the matched keypoints, which determines the pixel coordinates of the landmarks in different images. A map-building process is then performed to obtain the spatial positions of these landmarks. With the initial guesses of landmarks, bundle adjustment is used to optimize the camera parameters and the spatial positions of the landmarks. After the camera optimization, a geolocation method based on line-of-sight (LOS) is used to calculate the target geolocation based on the optimized camera parameters. The proposed method is validated through simulation and an experiment utilizing unmanned aerial vehicle (UAV) images, demonstrating its efficiency, robustness, and ability to achieve high-accuracy target geolocation.

A Recognition-Based Graphical Password System for Enhanced User Authentication

Abstract: Password-based authentication is a popular way for ensuring computer application security and privacy. Nonetheless, the "weakest link" in the authentication process is believed to be user-chosen weak passwords and hazardous input techniques. People generally utilize mnemonic or brief passphrases instead of random alphanumeric characters. Because internet and mobile applications are widely available, users can access them from any location, at any time, using any device. While this simplicity is generally appreciated, it also increases the likelihood that credentials will be compromised through Visual Hacking. The assault tactics can include employing external recording equipment or watching the victim personally in order to obtain their login information. In response to this issue, our team created a graphical password-based authentication mechanism to combat the threat of Visual Hacking. The system's security mechanism is three-tiered, featuring password verification as well as colour and pattern matching features. This means that even with numerous camera-based hacking attempts, potential attackers would have a tough time determining or limiting the password. In addition, we created and tested a system prototype to determine its usability. According to our trial results, our system can survive Visual Hacking better than previous authentication systems while maintaining a high level of security and usability.

Dense Feature Matching for Hazard Detection and Avoidance Using Machine Learning in Complex Unstructured Scenarios

Exploring the Moon and Mars are crucial steps in advancing space exploration. Numerous missions aim to land and research in various lunar locations, some of which possess challenging surfaces with unchanging features. Some of these areas are cataloged as lunar light plains. Their main characteristics are that they are almost featureless and reflect more light than other lunar surfaces. This poses a challenge during navigation and landing. This paper compares traditional feature matching techniques, specifically scale-invariant feature transform and the oriented FAST and rotated BRIEF, and novel machine learning approaches for dense feature matching in challenging, unstructured scenarios, focusing on lunar light plains. Traditional feature detection methods often need help in environments characterized by uniform terrain and unique lighting conditions, where unique, distinguishable features are rare. Our study addresses these challenges and underscores the robustness of machine learning. The methodology involves an experimental analysis using images that mimic lunar-like landscapes, representing these light plains, to generate and compare feature maps derived from traditional and learning-based methods. These maps are evaluated based on their density and accuracy, which are critical for effective structure-from-motion reconstruction commonly utilized in navigation for landing. The results demonstrate that machine learning techniques enhance feature detection and matching, providing more intricate representations of environments with sparse features. This improvement indicates a significant potential for machine learning to boost hazard detection and avoidance in space exploration and other complex applications.

An online long-term load forecasting method: Hierarchical highway network based on crisscross feature collaboration

In order to ensure the efficient operation of the power system, accurate online load forecasting is necessary. Existing studies face challenges in addressing unconventional load changes in long-term load forecasting (LTLF) and are unsuitable for online forecasting due to their reliance on measured Forecast Day Data (FDD). The development of a model independent of prior knowledge is essential to improve the interpretability and applicability of load forecasting methods. In this study, a hybrid model based on Crisscross Feature Collaboration (CFC) and Hierarchical Highway Network (HHN) is introduced. During the horizontal feature learning, the HHN is employed to learn temporal information from annual load curves across multiple time scales, reducing time complexity in LTLF. Subsequently, the prediction results of various time scales are combined to approximate FDD, addressing data leakage problems. In the vertical feature matching phase, the similar day data are screened by the multi-modal information of FDD and the Toeplitz Inverse Covariance Based Clustering (TICC) method, enabling the model to be independent of prior knowledge. Finally, through the HHN, all similar day data are utilized to reconstruct the forecasting load curve. In LTLF, the Root Mean Square Error (RMSE) of the proposed model reduces by 94.35 %, 85.74 %, 84.02 %, 90.72 %, 85.17 % and 87.76 % respectively in different scenarios compared to the direct forecasting. This signifies the effectiveness of our model in predicting the unconventional load change and achieving accurate online LTLF.

Feature Matching of Microsecond-Pulsed Magnetic Fields Combined with Fe3O4 Particles for Killing A375 Melanoma Cells.

The combination of magnetic fields and magnetic nanoparticles (MNPs) to kill cancer cells by magneto-mechanical force represents a novel therapy, offering advantages such as non-invasiveness, among others. Pulsed magnetic fields (PMFs) hold promise for application in this therapy due to advantages such as easily adjustable parameters; however, they suffer from the drawback of narrow pulse width. In order to fully exploit the potential of PMFs and MNPs in this therapy, while maximizing therapeutic efficacy within the constraints of the narrow pulse width, a feature-matching theory is proposed, encompassing the matching of three aspects: (1) MNP volume and critical volume of Brownian relaxation, (2) relaxation time and pulse width, and (3) MNP shape and the intermittence of PMF. In the theory, a microsecond-PMF generator was developed, and four kinds of MNPs were selected for in vitro cell experiments. The results demonstrate that the killing rate of the experimental group meeting the requirements of the theory is at least 18% higher than the control group. This validates the accuracy of our theory and provides valuable guidance for the further application of PMFs in this therapy.

An Underwater Localization Method Based on Visual SLAM for the Near-Bottom Environment

The feature matching of the near-bottom visual SLAM is influenced by underwater raised sediments, resulting in tracking loss. In this paper, the novel visual SLAM system is proposed in the underwater raised sediments environment. The underwater images are firstly classified based on the color recognition method by adding the weights of pixel location to reduce the interference of similar colors on the seabed. The improved adaptive median filter method is proposed to filter the classified images by using the mean value of the filter window border as the discriminant condition to retain the original features of the image. The filtered images are finally processed by the tracking module to obtain the trajectory of underwater vehicles and the seafloor maps. The datasets of seamount areas captured in the western Pacific Ocean are processed by the improved visual SLAM system. The keyframes, mapping points, and feature point matching pairs extracted from the improved visual SLAM system are improved by 5.2%, 11.2%, and 4.5% compared with that of the ORB-SLAM3 system, respectively. The improved visual SLAM system has the advantage of robustness to dynamic disturbances, which is of practical application in underwater vehicles operated in near-bottom areas such as seamounts and nodules.

Edge AI-Enabled Road Fixture Monitoring System

Effective monitoring of road fixtures is essential for urban safety and functionality. However, traditional inspections are time-consuming, costly, and error prone, while current automated solutions struggle with high initial setup costs, limited flexibility preventing wide adaptation, and reliance on centralized processing that can delay response times. This study introduces an edge AI-based remote road fixture monitoring system which automatically and continuously updates the information of the road digital twin (DT). The main component is a small-sized edge device consisting of a camera, GPS, and IMU sensors designed to be installed in typical cars. The device captures images, detects the fixture, and estimates their location by employing deep learning and feature matching. This information is transmitted to a dedicated cloud server and represented on a user-friendly user interface. Experiments were conducted to test the system’s performance. The results showed that the device could successfully detect the fixture and estimate their global coordinates. Outputs were marked and shown on the road DT, proving the integrated and smooth operation of the whole system. The proposed Edge AI device demonstrated that it could significantly reduce the data size by 80–84% compared to traditional methods. With a satisfactory object detection accuracy of 65%, the system effectively identifies traffic poles, stop signs, and streetlights, integrating these findings into a digital twin for real-time monitoring. The proposed system improves road monitoring by cutting down on maintenance and emergency response times, increasing the ease of data use, and offering a foundation for an overview of urban road fixtures’ current state. However, the system’s reliance on the quality of data collected under varying environmental conditions suggests potential improvements for consistent performance across diverse scenarios.

Критерии совмещения изображений в двухкамерной системе технического зрения

The image superposition is the important task in machine vision system. The scene which is shot by two-camera system is the same for each camera. So the images of each camera contain the same signal. This feature defines the possible criteria for image superposition. The research tests three criteria of superposition. These are criteria by correlation coefficient, the amount of the match contour points and the amount of match feature points. The numerical simulation was developed for this research. The results of simulation identified that the correlation coefficient criteria is the most noise resistance, but it demand the greatest processing time. The amount of match contour points criteria is the least noise resistance. The amount of match feature point’s criteria is better, than correlation coefficient criteria by noise resistance, and processing time is the least.

Analysis of Music Brand Similarity Levels Using a Visual Computing Approach

Currently, many trademarks have emerged, especially for music equipment products circulating in the community. Especially in Indonesia, musical instrument brands with various models have emerged. The problem occurs when the brand is considered to resemble the original brand, which makes it uncomfortable for musicians. Even though the quality of the tone produced is not assessed from a brand perspective, it is felt by brand owners to be quite a violation of the code of ethics. This has an impact on marketing products that are considered genuine. In this paper, a concept will be proposed in determining whether a trademark is considered authentic in terms of the logo. Use of Visual Computing with SIFT and SURF algorithms. SIFT (Scale-Invariant Feature Transform) and SURF (Speeded-Up Robust Features) are two popular algorithms for feature extraction and matching in image processing. Both are frequently used in computer vision applications such as object detection, image matching, and object recognition. The results of this analysis will be used for musicians who want to buy musical equipment to be able to detect the logo to compare the success of a musical instrument product. The appropriateness of the quality itself will be discussed in a different discussion. The benefits that can be generated will make musicians, especially in Indonesia, more conscientious. The possibility of sorting so that you can like your own product will be very possible.

CattNIS: Novel identification system of cattle with retinal images based on feature matching method

Small and large farm animals are frequently identified using techniques like microchipping or tagging. However, these techniques may have a detrimental impact on animal welfare and may result in theft and fraud. Modern and sophisticated identification procedures using biometrics, including retina, face, and iris recognition, are available and have less of a detrimental effect on animal welfare. Because biometric markers are unchangeable and unaffected, it is feasible to stop situations of theft and fraud. This study uses retina images to identify animals based on each animal's different and specific retinal patterns. To the best of our knowledge, there isn't a publicly accessible dataset for images of animal retinas. The three main goals of this study are to: 1) produce a novel dataset of Cattle retinal images and make it publicly available; 2) create a successful image processing-based system for retinal identification and recognition; and 3) create a graphical user interface (GUI) that enables querying the database to ascertain whether a specific Cattle retina is present. To accomplish these goals, a dataset comprising 2430 images from the left and right eyes of 300 cattle was produced. After the dataset was created, the animal retinal images were segmented using image processing methods such as scaling, color transformations, image sharpening, contrast enhancements, noise filtering, and histogram equalization. By using the segmented images with the SURF, FAST, BRISK, and HARRIS techniques, distinctive retinal features were found. During the identification phase, the animal whose retinal characteristic matched the sought animal's retina the closest was recognized. Within the developed CattNIS system, the SURF approach has the highest accuracy rate of 92.25 percent. The results of this study show that the identification of cattle using retinal vascular patterns is highly successful. The availability of the collected cattle retinal images from this study is expected to significantly contribute to the advancement of future research in this field, both in terms of quantity and quality.

A stereovision-based efficient measurement approach for surface flatness of concrete members

The surface flatness of concrete members is an important quality characteristic of buildings and civil infrastructure. Existing measurement methods such as the straightedge method, F-numbers method, and laser scanning method have various limitations such as inefficiency in engineering practices, low accuracy, and high costs. Therefore, this paper presents a novel stereovision-based system comprising two industrial cameras and one projector alongside optimized algorithms to efficiently measure the surface flatness of concrete member. The proposed system is highlighted by effective feature point extraction and matching algorithms, the innovative introduction of feature point extraction error to system measurement error evaluation, and enriched flatness indicators to comprehensively assess surface flatness of concrete members. The outputs of the system, namely, the contour maps as visual demonstrations and the flatness indicator results, can collectively reflect various roughness conditions with high speed and precision. Experiments at multiple levels have validated the proposed algorithms and measurement system, showing that it is capable for providing fast and accurate measures for the surface flatness of concrete members. Further research and developments are sought in system optimization and expanding its applications to other structural members and engineering practices.

Improving diversity of speech‐driven gesture generation with memory networks as dynamic dictionaries

AbstractGenerating co‐speech gestures for interactive digital humans remains challenging because of the indeterministic nature of the problem. The authors observe that gestures generated from speech audio or text by existing neural methods often contain less movement shift than expected, which can be viewed as slow or dull. Thus, a new generative model coupled with memory networks as dynamic dictionaries for speech‐driven gesture generation with improved diversity is proposed. More specifically, the dictionary network dynamically stores connections between text and pose features in a list of key‐value pairs as the memory for the pose generation network to look up; the pose generation network then merges the matching pose features and input audio features for generating the final pose sequences. To make the improvements more accurately measurable, a new objective evaluation metric for gesture diversity that can remove the influence of low‐quality motions is also proposed and tested. Quantitative and qualitative experiments demonstrate that the proposed architecture succeeds in generating gestures with improved diversity.

SADGFeat: Learning local features with layer spatial attention and domain generalization

Local feature plays a pivotal role in various robotic tasks, including 3D reconstruction and visual localization. Although deep learning-based local features have proven superior to their traditional counterparts, they still face challenges in practical applications due to matching failures. These challenges primarily stem from inaccuracies in keypoint localization and the limited robustness of descriptors particularly with significant appearance changes. In this study, we introduce a novel method for local feature learning based on layer spatial attention and domain generalization. Firstly, a keypoint extraction strategy driven by layer spatial attention from high-level feature is proposed to enhance the accuracy of keypoint localization, progressing from coarse to fine granularity. Secondly, a new learning paradigm based on domain generalization is developed to extract local features with resilience against variations in illumination. To enrich the domain diversity within the training dataset, a real-time and lossless Fourier transform-based domain augmentation method is introduced. This method seamlessly integrates into the training process, enhancing the model's adaptability to varying domains. Additionally, explicit feature alignment-based representation learning is performed, further reinforcing the extraction of domain-invariant local features. Experimental results on public datasets demonstrate that the proposed method achieves state-of-the-art performance across various downstream tasks reliant on local feature matching, such as image matching, 3D reconstruction, and long-term visual localization.

Chinese-Vietnamese cross-lingual event retrieval method based on knowledge distillation

Cross-lingual event retrieval is an information retrieval task aimed at cross-lingual event retrieval among multiple languages to find text or documents related to a specific event. Specific to Chinese-Vietnamese cross-language event retrieval, it involves using Chinese as a query to retrieve Vietnamese documents related to the query event. The critical issue is how to efficiently align query and document representations with limited resources. Existing cross-language pre-training models are trained on large-scale multilingual corpora, but their training goals do not include explicit language alignment tasks. Due to the uneven distribution of training corpora between different languages, these models have The problem of language bias. Therefore, this linguistic bias is also inherited in cross-lingual retrieval based on these models. To solve this problem, this paper proposes a Chinese-Vietnamese cross-lingual event retrieval method based on knowledge distillation. This approach enables the model to learn good query-document matching features from monolingual retrieval by transferring knowledge from high-resource to low-resource languages. By enhancing the alignment between queries and documents in different languages in a shared semantic space, the method improves the performance of Chinese-Vietnamese cross-lingual event retrieval.

Research on Full-Field Dynamic Deflection Measurement of Beams Based on Dense Feature Matching and Mismatch Removal Method

To solve the problems of measurement errors led by mismatches of dense feature matching in machine vision structural deflection measurement, this paper proposes a dense feature extraction, matching, and dual-step mismatch-removal-based full-field structural dynamic deflection measurement method. First, the of dense feature detection and matching theory is introduced to extract the SIFT feature points on a structural surface in an image sequence and matched by FLANN to trace the structure movement, and the mechanisms and causes of mismatches are analyzed. Then, a dual-step mismatch removal method combining RANSAC and Structural Displacement Continuity Restriction (SDCR) is introduced to achieve full-field dynamic beam deflection measurement. The proposed method is validated through indoor cantilever beam experiments, and results show that the method can effectively eliminate a large number of SIFT feature mismatches (accounting for approximately 55% of the total matched feature points). The full-field dynamic displacement field of the beam can be measured with the correctly matched dense feature points by converting dense feature point displacements into continuous and uniform spatiotemporal deflections of the structure. A comparison with the GOM Correlate Professional DIC measurement system was conducted, and the maximum measurement error of the cantilever beam dynamic displacement of the proposed method is between 0.024 and 0.053 mm, the root mean squared error of displacement is approximately 0.01 mm, and the correlation coefficient between two deflection–time curves reaches 0.9964. The proposed algorithm is proven to be effective in full-field displacement measurement and has great potential in future structural health monitoring of bridges.

Augmented reality presentation system of skeleton image based on biomedical features

Aimed at limitations in the description and expression of three-dimensional (3D) physical information in two-dimentsional (2D) medical images, feature extraction and matching method based on the biomedical characteristics of skeletons is employed in this paper to map the 2D images of skeletons into a 3D digital model. Augmented reality technique is used to realize the interactive presentation of skeleton models. Main contents of this paper include: Firstly, a three-step reconstruction method is used to process the bone CT image data to obtain its three-dimensional surface model, and the corresponding 2D–3D bone library is established based on the identification index of the 2D image and the 3D model; then, a fast and accurate feature extraction and matching algorithm is developed to realize the recognition, extraction, and matching of 2D skeletal features, and determine the corresponding 3D skeleton model according to the matching result. Finally, based on the augmented reality technique, an interactive immersive presentation system is designed to achieve visual effects of the virtual human bone model superimposed and rendered in the world scenes, which improves the effectiveness of information expression and transmission, as well as the user's immersion and embodied experience.