Feature Vector Of Image Research Articles

The analysis of art design under improved convolutional neural network based on the Internet of Things technology

This work aims to explore the application of an improved convolutional neural network (CNN) combined with Internet of Things (IoT) technology in art design education and teaching. The development of IoT technology has created new opportunities for art design education, while deep learning and improved CNN models can provide more accurate and effective tools for image processing and analysis. In order to enhance the effectiveness of art design teaching and students’ creative expression, this work proposes an improved CNN model. In model construction, it increases the number of convolutional layers and neurons, and incorporates the batch normalization layer and dropout layer to enhance feature extraction capabilities and reduce overfitting. Besides, this work creates an experimental environment using IoT technology, capturing art image samples and environmental data using cameras, sensors, and other devices. In the model application phase, image samples undergo preprocessing and are input into the CNN for feature extraction. Sensor data are concatenated with image feature vectors and input into the fully connected layers to comprehensively understand the artwork. Finally, this work trains the model using techniques such as cross-entropy loss functions and L2 regularization and adjusts hyperparameters to optimize model performance. The results indicate that the improved CNN model can effectively acquire art sample data and student creative expression data, providing accurate and timely feedback and guidance for art design education and teaching, with promising applications. This work offers new insights and methods for the development of art design education.

A new adversarial malware detection method based on enhanced lightweight neural network

With the gradual expansion of Android systems from mobile phones to intelligent devices, a huge amount of malware has been found every year. To improve the malware detection performance and reduce its reliance on expert experience, deep learning technology has been widely used. However, as the complexity of deep learning models continues to increase, it rapidly increases the consumption of hardware resources. At the same time, anti-detection technology such as Generative Adversarial Networks (GANs) are widely used to evade Artificial Intelligence (AI)-based detection methods. In this paper, we propose a new classification model based on an improved lightweight neural network that can effectively improve the execution efficiency and detection performance of malware detection methods against adversarial malware samples. First, our method uses local-information-entropy-based image generation technology to construct effective image feature vectors. Then, the performance of the lightweight neural network model ESPNetV2 is improved from four aspects. Finally, a new adversarial malware generation model called Mal-WGANGP is proposed. It can automatically generate a large number of adversarial samples to robust our model. In order to evaluate our method, we construct several experiments and compare the detection performance of our method with 19 other novel efficient neural network detection models. Experimental results show that our image enhancement method and detection model have the highest detection accuracy of adversarial samples.

Image retrieval based on auto-encoder and clustering with centroid update

Abstract This paper suggests a revolutionary deep learning method using a dynamic deep auto-encoder for improving the performance of indexing the feature vectors of images by centroid updation. Feature vectors such as color, semantic, and spatial local binary pattern are extracted from the images for content-based image retrieval. The owner encrypts the images for protection using elliptic curve cryptography before uploading them to the cloud. A black hole entropic fuzzy clustering with Tversky indexing is used to retrieve similar information. When the new training image is matched with any of the centroid then the centroid gets updated by using dynamic deep auto-encoder. During the auto-encoder phase, the conflicted data points are dedicated to reconstruction and the reliable data points are helpful to centroid updation. The suggested BHE fuzzy clustering with dynamic deep auto-encoder approach fared better than the current methods, achieving the best accuracy of 97.605%, the highest $\boldsymbol{F_{1}}$ score of 90.210%, better precision of 90.001%, and the highest recall of 95.149%.

Sensitivity Study of Feature Vectors Formed on the Basis of Multiple Scale Transformations of Processed Images

Introduction. Under the conditions of globalization, transportation of oil products by tanker fleets becomes one of the causes of man-made disasters in the water areas of seas and oceans. In this context, environmental monitoring acquires particular significance as a tool for ensuring timely detection of negative consequences of man-made disasters. This task is facilitated by recognition of images obtained from unmanned aerial vehicles with selection of those depicting the traces of man-made accidents or their consequences.Aim. To develop approaches for carrying out automatic selection of input data obtained from unmanned aerial vehicles in the form of photo and video images at the preliminary stage of image recognition.Materials and methods. The theoretical part of the study employed a classification method based on pattern recognition theory. Mathematical processing and calculation were carried out in the MATLAB environment. Simulation was conducted using the MathCAD environment.Results. A series of experiments was conducted to select a basis for discrete wavelet transforms. Modeling was conducted using the study results of the sensitivity of the feature vectors of images formed on the basis of different types of wavelet transforms. A concept of image construction for the purposes of feature vector formation was developed.Conclusion. An approach to the formalization of images for the purposes of feature vector formation is proposed. A metric of their contrast estimation is substantiated. It was established that the sensitivity of a recognition system based on the representation of images in the form of discrete wavelet transform matrices depends not only on the type of the mother wavelet, but also on the value of the scale parameter.

Aligning the domains in cross domain model inversion attack

Model Inversion Attack reconstructs confidential training dataset from a target deep learning model. Most of the existing methods assume the adversary has an auxiliary dataset that has similar distribution with the private dataset. However, this assumption does not always hold in real-world scenarios. Since the private dataset is unknown, the domain divergence between the auxiliary dataset and the private dataset is inevitable. In this paper, we use Cross Domain Model Inversion Attack to represent the distribution divergence scenario in MIA. With the distribution divergence between the private images and auxiliary images, the distribution between the feature vectors of the private images and those of the auxiliary images is also different. Moreover, the outputted prediction vectors of the auxiliary images are also misclassified. The inversion attack is thus hard to be performed. We perform both the feature vector inversion task and prediction vector inversion task in this cross domain setting. For feature vector inversion, Domain Alignment MIA (DA-MIA) is proposed. While performing the reconstruction task, DA-MIA aligns the feature vectors of auxiliary images with the feature vectors of private images in an adversarial manner to mitigate the domain divergence between them. Thus, semantically meaningful images can be reconstructed. For prediction vector inversion, we further introduce an auxiliary classifier and propose Domain Alignment MIA with Auxiliary Classifier (DA-MIA-AC). The auxiliary classifier is pretrained by the auxiliary dataset and fine-tuned during the adversarial training stage. Thus, the misclassification problem caused by domain divergence can be solved, and the images can be reconstructed correctly. Various experiments are performed to show the advancement of our methods, the results show that DA-MIA can improve the SSIM score of the reconstructed images for up to 191%, DA-MIA-AC can increase the classification accuracy score of the reconstructed images from 9.18% to 81.32% in Cross Domain Model Inversion Attack.

The use of pretrained neural networks for solving the problem of reverse searching of X-ray images of prohibited items and substances

The paper considers the application of pretrained neural networks to solve the problem of reverse searching of X-ray images of prohibited items and substances. The purpose of the work is to conduct an analysis and substantiate ways to improve the efficiency of baggage and passenger hand luggage X-ray image recognition systems. An analysis of existing domestic and foreign works in the field of baggage and passenger hand luggage X-ray image recognition is presented. It has been revealed that, despite the achieved results in the development of algorithms for recognizing prohibited items and substances, they do not fully cope with such a complexity factor as the overlay of objects. To solve this problem, the paper proposes to additionally analyze X-ray images with low confidence in object recognition. This stage includes the following steps: image segmentation, extraction of features of segmented image elements; search for similar images in the database; decision-making on the class of segmented image elements. This article discusses the last three steps. Variants of approaches to feature extraction from images are analyzed, particularly those based on the application of convolutional autoencoders and pretrained neural networks. The approach based on the application of pretrained neural networks is chosen. The ResNet-50 architecture neural network, pretrained on the ImageNet collection, is used during the work. In order to apply this model to extract image feature vectors, the last classification layer was preliminarily removed. All the previous layers of the model encode the image into a vector. ResNet-50 generates a 2048-dimensional feature vector of images. The principal component analysis is used to reduce the dimensionality of the image feature vectors. The decision of whether the segmented image element is a prohibited item or substance is considered as a reverse search problem using the k-nearest neighbor algorithm. In this case, the class of the X-ray image element is the class most frequently encountered among the K nearest neighbors. In order to test the proposed approach, a training dataset, including 4,635 images of individual items and substances that may be encountered in baggage and passenger hand luggage, was generated. A comparative analysis of image indexing and image search under different algorithms and feature number is presented. A comparative analysis of the model accuracy is provided. It is concluded that the most acceptable is the “Brute force” algorithm in combination with the principal component analysis.

Unifying Multi-Modal Uncertainty Modeling and Semantic Alignment for Text-to-Image Person Re-identification

Text-to-Image person re-identification (TI-ReID) aims to retrieve the images of target identity according to the given textual description. The existing methods in TI-ReID focus on aligning the visual and textual modalities through contrastive feature alignment or reconstructive masked language modeling (MLM). However, these methods parameterize the image/text instances as deterministic embeddings and do not explicitly consider the inherent uncertainty in pedestrian images and their textual descriptions, leading to limited image-text relationship expression and semantic alignment. To address the above problem, in this paper, we propose a novel method that unifies multi-modal uncertainty modeling and semantic alignment for TI-ReID. Specifically, we model the image and textual feature vectors of pedestrian as Gaussian distributions, where the multi-granularity uncertainty of the distribution is estimated by incorporating batch-level and identity-level feature variances for each modality. The multi-modal uncertainty modeling acts as a feature augmentation and provides richer image-text semantic relationship. Then we present a bi-directional cross-modal circle loss to more effectively align the probabilistic features between image and text in a self-paced manner. To further promote more comprehensive image-text semantic alignment, we design a task that complements the masked language modeling, focusing on the cross-modality semantic recovery of global masked token after cross-modal interaction. Extensive experiments conducted on three TI-ReID datasets highlight the effectiveness and superiority of our method over state-of-the-arts.

DIC-Transformer: interpretation of plant disease classification results using image caption generation technology

Disease image classification systems play a crucial role in identifying disease categories in the field of agricultural diseases. However, current plant disease image classification methods can only predict the disease category and do not offer explanations for the characteristics of the predicted disease images. Due to the current situation, this paper employed image description generation technology to produce distinct descriptions for different plant disease categories. A two-stage model called DIC-Transformer, which encompasses three tasks (detection, interpretation, and classification), was proposed. In the first stage, Faster R-CNN was utilized to detect the diseased area and generate the feature vector of the diseased image, with the Swin Transformer as the backbone. In the second stage, the model utilized the Transformer to generate image captions. It then generated the image feature vector, which is weighted by text features, to improve the performance of image classification in the subsequent classification decoder. Additionally, a dataset containing text and visualizations for agricultural diseases (ADCG-18) was compiled. The dataset contains images of 18 diseases and descriptive information about their characteristics. Then, using the ADCG-18, the DIC-Transformer was compared to 11 existing classical caption generation methods and 10 image classification models. The evaluation indicators for captions include Bleu1–4, CiderD, and Rouge. The values of BLEU-1, CIDEr-D, and ROUGE were 0.756, 450.51, and 0.721. The results of DIC-Transformer were 0.01, 29.55, and 0.014 higher than those of the highest-performing comparison model, Fc. The classification evaluation metrics include accuracy, recall, and F1 score, with accuracy at 0.854, recall at 0.854, and F1 score at 0.853. The results of DIC-Transformer were 0.024, 0.078, and 0.075 higher than those of the highest-performing comparison model, MobileNetV2. The results indicate that the DIC-Transformer outperforms other comparison models in classification and caption generation.

An Analysis of the Application of Motion Capture Technology in the Creation of Contemporary Dance with the Integration of Fibonacci Sequences

Abstract This paper combines the existing development of Fibonacci sequence aesthetics in art creation for the aesthetic planning of the Fibonacci sequence on music rhythm, proposes to design a new modern dance creation method by combining the diversified rotation methods of the Fibonacci sequence, and builds a process framework for generating Fibonacci sequence dance score based on motion capture. The wavelet transform is employed to determine the start of the dance movement, and the gradient descent algorithm is utilized to analyze the time coordinates and refine the dance movement. Recognize the segmented elemental action sequences using the DTW method, propose three basic actions from the dataset, and test the recognition rate of the DTW algorithm for each action. Generate a Fibonacci sequence dance score by combining human gestures. Combined with the research scoring, analyze the artistic talent of the Fibonacci sequence-generated dance. According to the data results, it can be seen that the action feature vector of the two-handed waving image is P=[0.8968, 0.9925, 0.9124, 1.0243, 1.0198, 1.2438, 1.4596, 1.3698, 1.5868, 1.6596, 1.5029] with 98.67% of the two-handed waving recognition rate (with a standard deviation of 0.25). The evaluation of Fibonacci sequence generation dance by research users tends to be moderate.

CMFAN: Cross-Modal Feature Alignment Network for Few-Shot Single-View 3D Reconstruction.

Few-shot single-view 3D reconstruction learns to reconstruct the novel category objects based on a query image and a few support shapes. However, since the query image and the support shapes are of different modalities, there is an inherent feature misalignment problem damaging the reconstruction. Previous works in the literature do not consider this problem. To this end, we propose the cross-modal feature alignment network (CMFAN) with two novel techniques. One is a strategy for model pretraining, namely, cross-modal contrastive learning (CMCL), here the 2D images and 3D shapes of the same objects compose the positives, and those from different objects form the negatives. With CMCL, the model learns to embed the 2D and 3D modalities of the same object into a tight area in the feature space and push away those from different objects, thus effectively aligning the global cross-modal features. The other is cross-modal feature fusion (CMFF), which further aligns and fuses the local features. Specifically, it first re-represents the local features with the cross-attention operation, making the local features share more information. Then, CMFF generates a descriptor for the support features and attaches it to each local feature vector of the query image with dense concatenation. Moreover, CMFF can be applied to multilevel local features and brings further advantages. We conduct extensive experiments to evaluate the effectiveness of our designs, and CMFAN sets new state-of-the-art performance in all of the 1-/10-/25-shot tasks of ShapeNet and ModelNet datasets.

Japanese historical character recognition by focusing on character parts

Japanese historical documents provide valuable information. Character recognition is a critical technology for the digitalization of historical documents. Sample imbalance is a significant obstacle in recognizing Japanese historical characters, kuzushiji. Thousands of kuzushiji only have less than a few samples. Thus, recognition performance deteriorates greatly in kuzushiji with a few samples. In this study, we propose a framework for transferring knowledge of character parts from font to kuzushiji. The pretraining learns character parts from synthesized font images. However, fine-tuning to kuzushiji is more complex. We propose calculating a mean squared error loss between feature vectors of kuzushiji and font images, resulting in consistent feature vectors in kuzushiji and font. Consequently, we can perform zero-shot recognition for kuzushiji using the font images of zero-sampled kuzushiji. The experimental results show that the proposed method recognized zero-sampled kuzushiji at approximately 48% accuracy. Consequently, we significantly expand the number of recognizable kuzushiji.

Hybrid watermarking algorithm for medical images based on digital transformation and MobileNetV2

The current study presents a new methodology for zero watermarking that relies on the discrete wavelet transform (DWT) and an improved MobileNetV2 convolutional neural network, along with the discrete cosine transform (DCT). This model is intended to overcome the issue of algorithm robustness in encrypted medical image watermarking. The proposed technique targets medical images inside the area of encryption and offers a unique watermarking approach to overcome the issues above. The coefficients acquired from the fully linked network layer are converted using DWT and DCT to create the medical image's feature vector. Second, MobileNetV2 is initially fed the medical image; this network has been prepared by adjusting parameters such as the convolution kernels' size and the convolution modules' typical architecture. Finally, the encryption of watermarks is achieved through the utilization of the logistic map system and hash function, whereby an independent party securely stores the requisite keys. The integration of a zero-watermark involves the execution of logical operations on both the encrypted watermarks and the attributes of the source image. The results of the experiment indicate that the algorithm can effectively differentiate encrypted medical images and recover the initial data from encrypted watermarked material despite conventional and geometric attacks. Compared to alternative algorithms, their superior resilience and invisibility are noteworthy.

GAGPT-2: A Geometric Attention-based GPT-2 Framework for Image Captioning in Hindi

Image captioning frameworks usually employ an encoder-decoder paradigm, with the encoder receiving abstract image feature vectors as input and the decoder for language modeling. Nowadays, most prominent architectures employ features from region proposals derived from object detection modules. In this work, we propose a novel architecture for image captioning. We employ the object detection module integrated with transformer architecture as an encoder and GPT-2 (Generative Pre-trained Transformer) as a decoder. The encoder utilizes the information of the spatial relationships among detected objects. We introduce a unique methodology for image caption generation in Hindi, which is widely spoken in South Asia and India and is the world’s third most spoken language as well as India’s official language. In terms of BLEU scores, the proposed approach’s performance is comparable to those of other baselines, and the results illustrate that the proposed approach outperforms the other baselines. The efficacy of the proposed approach in generating correct captions is further determined by human assessment in terms of adequacy and fluency.

Deep multitask learning with label interdependency distillation for multicriteria street-level image classification

Multitask learning (MTL) aims at beneficial joint solving of multiple prediction problems by sharing information across different tasks. However, without adequate consideration of interdependencies, MTL models are prone to miss valuable information. In this paper, we introduce a novel deep MTL architecture that specifically encodes cross-task interdependencies within the setting of multiple image classification problems. Based on task-wise interim class label probability predictions by an intermediately supervised hard parameter sharing convolutional neural network, interdependencies are inferred in two ways: i) by directly stacking label probability sequences to the image feature vector (i.e., multitask stacking), and ii) by passing probability sequences to gated recurrent unit-based recurrent neural networks to explicitly learn cross-task interdependency representations and stacking those to the image feature vector (i.e., interdependency representation learning). The proposed MTL architecture is applied as a tool for generic multi-criteria building characterization using street-level imagery related to risk assessments toward multiple natural hazards. Experimental results for classifying buildings according to five vulnerability-related target variables (i.e., five learning tasks), namely height, lateral load-resisting system material, seismic building structural type, roof shape, and block position are obtained for the Chilean capital Santiago de Chile. Our MTL methods with cross-task label interdependency modeling consistently outperform single task learning (STL) and classical hard parameter sharing MTL alike. Even when starting already from high classification accuracy levels, estimated generalization capabilities can be further improved by considerable margins of accumulated task-specific residuals beyond +6% κ. Thereby, the combination of multitask stacking and interdependency representation learning attains the highest accuracy estimates for the addressed task and data setting (up to cross-task accuracy mean values of 88.43% overall accuracy and 84.49% κ). From an efficiency perspective, the proposed MTL methods turn out to be substantially favorable compared to STL in terms of training time consumption.

Novel in-house knowledge-based automated planning system for lung cancer treated with intensity-modulated radiotherapy.

The goal of this study was to propose aknowledge-based planning system which could automatically design plans for lung cancer patients treated with intensity-modulated radiotherapy (IMRT). From May 2018 to June 2020, 612 IMRT treatment plans of lung cancer patients were retrospectively selected to construct aplanning database. Knowledge-based planning (KBP) architecture named αDiar was proposed in this study. It consisted of two parts separated by afirewall. One was the in-hospital workstation, and the other was the search engine in the cloud. Based on our previous study, A‑Net in the in-hospital workstation was used to generate predicted virtual dose images. Asearch engine including athree-dimensional convolutional neural network (3D CNN) was constructed to derive the feature vectors of dose images. By comparing the similarity of the features between virtual dose images and the clinical dose images in the database, the most similar feature was found. The optimization parameters (OPs) of the treatment plan corresponding to the most similar feature were assigned to the new plan, and the design of anew treatment plan was automatically completed. After αDiar was developed, we performed two studies. The first retrospective study was conducted to validate whether this architecture was qualified for clinical practice and involved 96patients. The second comparative study was performed to investigate whether αDiar could assist dosimetrists in improving the quality of planning for the patients. Two dosimetrists were involved and designed plans for only one trial with and without αDiar; 26patients were involved in this study. The first study showed that about 54% (52/96) of the automatically generated plans would achieve the dosimetric constraints of the Radiation Therapy Oncology Group (RTOG) and about 93% (89/96) of the automatically generated plans would achieve the dosimetric constraints of the National Comprehensive Cancer Network (NCCN). The second study showed that the quality of treatment planning designed by junior dosimetrists was improved with the help of αDiar. Our results showed that αDiar was an effective tool to improve planning quality. Over half of the patients' plans could be designed automatically. For the remaining patients, although the automatically designed plans did not fully meet the clinical requirements, their quality was also better than that of manual plans.

De-identifying transmission system using wireless channel as differential privacy noise and deep neural networks

In order to deal with the unprecedented problem of data privacy in deep-learning based computer vision technologies, we propose a de-identifying transmission system that utilizes a wireless channel as differential privacy (DP) noise and is constructed using deep neural networks. By following the Gaussian mechanism of DP, we present that the signal received at a receiver can be considered as a Gaussian mechanism. Then, we discuss the relationship between the signal-to-noise ratio of a received signal and the privacy budget of DP, and introduce how to control transmit power to achieve a specific privacy budget. The proposed system can be divided into three parts, named transmitter, wireless channel, and receiver. The transmitter and receiver are constructed by using deep-learning networks. The transmitter consists of an image encoder network based on a neural network and a power control module that controls the transmit power for adjusting the level of de-identification. The wireless channel acts as differential privacy noise, and anonymizes the transmitted image feature vector extracted from the transmitter’s image encoder network. The receiver includes a post-processing network and an image decoder network, which are also implemented by using deep neural networks. The post-processing network is proposed for high-quality decoded face images at the receiver, which maps received feature vectors perturbed by a wireless channel back into the latent space of deep-learning based image encoder and decoder networks. Finally, extensive qualitative and quantitative evaluations confirm that the proposed system can well de-identify transmitted face images only by controlling the transmit power while maintaining the usefulness of decoded face images. The proposed system shows that the Recall@1 is smaller than 5.2 and the face detection probability is larger than 90 % at SNR=2 dB. Since the de-identification process is performed by wireless channel noise, the proposed system does not require de-identification processing at a transmitter, and thus there is no burden on the transmitter to anonymize face images. Moreover, the additional advantage of this proposed system is that the level of de-identification can be controlled only by changing the transmit power. This proposed de-identification system can be utilized in a wireless image acquisition scenario where images are captured from a wireless edge, such as a CCTV camera, and then sent to a server while protecting people’s privacy.

Image Retrieval Using Fusion of Sauvola and Thepade’s Sorted Block Truncation Coding-Based Color Features

Because of the tremendous growth in digital imaging, enhanced communication and storage technology, billions of images are captured, stored, and exchanged daily. Finding and searching for an image in a large collection is becoming challenging. The query by reference image retrieval (IR) technique aims to close the semantic gap between the query and retrieve images while improving performance. The primary goal of the work proposed here is to develop discriminative and descriptive features of the image with the minimum possible size. Here, the weighted feature fusion-based IR technique is proposed using Sauvola local thresholding (SLT) and Thepade’s Sorted Block Truncation Coding (SBTC) methods. The proposed technique is tested using two standard datasets with mean square error (MSE) as a distance measure and average retrieval accuracy (ARA) as a performance metric. The technique has contributed to the enhancement of ARA with the small and fixed-size image feature vector. The feature vector generated is much smaller than the image dimension and is used as a feature vector to represent the image for retrieval. Results prove that the proposed technique of SBTC 8-ary with 0.1 weight and SLT with 0.9 weight feature fusion gives better ARA than other techniques studied.

T2-fuzzy multi-fused facial image fusion (T2FMFImgF): An efficient face recognition

This paper presents a novel decision-making method for face recognition where the features were extracted from the original image fused with its corresponding true and partial diagonal images. To extract features, we adopted the generalized two-dimensional FLD (G2DFLD) feature extraction technique. The feature vectors from a test image are given as input to neural network-based classifier. It is trained with the feature vectors of original image and diagonally fused images and thereby the merit weights with respect to different classes were generated. To address the factors that affect the face recognition accuracy and uncertainty related to raw biometric data, a fuzzy score for each of the classes is generated by treating a type-2 fuzzy set. This type-2 fuzzy set is formed by the feature vectors of both the diagonally fused training samples and the test image of the respective classes. A concluding score for each of the classes under consideration is computed by fusing complemented merit weight with the complemented fuzzy score. These class-wise concluding scores are considered in the face recognition process. In this study, the well-known face databases (AT&T, UMIST and CMU-PIE) are used to evaluate the performance of the proposed method. The experimental results illustrate the fact that the proposed method has exhibited superior classification precision as compared with other state-of-art methods. Our T2FMFImgF method achieves highest face recognition accuracies of 99.41%, 98.36% and 89.80% in case of AT&T, UMIST and CMU-PIE (with expression), respectively while for CMU-PIE (with Light) the highest recognition accuracy is 97.957%. In addition to it, the presented method is quite successful in fusing and classifying textural information from the original and partial diagonal images by integrating them with type-2 fuzzy set-based treatment.

An encoded histogram of ridge bifurcations and contours for fingerprint presentation attack detection

In recent years, the exponential growth of internet technologies has made personal authentication an integral part of security applications. The fingerprint-based biometric systems are essentially used to safeguard the users' privacy and confidentiality. However, such systems are prone to spoof attacks by artificial replicas of the fingerprints. This paper presents an improved feature extractor called BiRi-PAD (Encoded Histogram of Ridge Bifurcations and Contours for fingerprint Presentation Attack Detection) that aims to enhance the accuracy of live fingerprint detection. The feature extraction process of the proposed BiRi-PAD consists of four steps. First, the fingerprint image undergoes a process of extracting 2-channel ridge contour maps (2-RC maps). The first channel of 2-RC maps consists of ridge contours extracted by a set of derivative filters in the spatial domain whereas the second channel of 2-RC maps consists of the ridge contours extracted by the maximum moments based on phase congruency in the frequency domain. Further, minutiae-based feature information i.e., ridge bifurcations are extracted by the minimum moments based on phase congruency covariance. Moreover, a fusion equation is proposed to integrate 2-RC maps and bifurcations into a single feature map. Second, an improved Comprehensive Local Phase Quantization (CLPQ) based on the well-known feature descriptor Rotation Invariant Local Phase Quantization (LPQri) is proposed to extract the phase information of ridges. CLPQ extracts the orientation of the ridges by using the complex parts of the significant frequency components of LPQri and monogenic filters. Third, the proposed BiRi-PAD quantizes the 2-RC maps into pre-determined intervals. Finally, both 2-RC maps and CLPQ features are integrated to generate a feature vector of a single fingerprint image. Performance evaluations of BiRi-PAD are conducted on three publicly available benchmarks from the LivDet competition, namely LivDet 2013, 2011, and 2015. Experimental evaluations demonstrate that the proposed BiRi-PAD achieves significant reductions in average rates compared to state-of-the-art techniques of fingerprint liveness detection. Specifically, on LivDet 2013, LivDet 2011, and LivDet 2015, the average rates are reduced to 1.92%, 4.39%, and 4.55%, respectively.

Building Change Detection in Remote Sensing Imagery with Focal Self-Attention and Multi-Level Feature Fusion

Accurate and intelligent building change detection greatly contributes to effective urban development, optimized resource management, and informed decision-making in domains such as urban planning, land management, and environmental monitoring. Existing methodologies face challenges in effectively integrating local and global features for accurate building change detection. To address these challenges, we propose a novel method that uses focal self-attention to process the feature vector of input images, which uses a “focusing” mechanism to guide the calculation of the self-attention mechanism. By focusing more on critical areas when processing image features in different regions, focal self-attention can better handle both local and global information, and is more flexible and adaptive than other methods, improving detection accuracy. In addition, our multi-level feature fusion module groups the features and then constructs a hierarchical residual structure to fuse the grouped features. On the LEVIR-CD and WHU-CD datasets, our proposed method achieved F1-scores of 91.62% and 89.45%, respectively. Compared with existing methods, ours performed better on building change detection tasks. Our method therefore provides a framework for solving problems related to building change detection, with some reference value and guiding significance.