Image Feature Extraction Research Articles

Quantum computational infusion in extreme learning machines for early multi-cancer detection

A timely and accurate cancer diagnosis is essential for improving treatment outcomes. This study presents a hybrid model integrating Extreme Learning Machine (ELM) with FuNet transfer learning, applied on a multi-cancer dataset and optimized using the Quantum-Genetic Binary Grey Wolf Optimizer (Q-GBGWO). This model leverages a diverse feature fusion strategy, enhancing the extraction of critical imaging features, while Q-GBGWO optimizes ELM parameters to achieve superior classification performance. Results demonstrate that Q-GBGWO-ELM improves diagnostic accuracy by an average of 6.5% compared to traditional methods, with notable accuracy rates across various cancers: 98.80% for breast cancer, 92.30% for brain tumors, 97.00% for skin cancer, and 96.98% for lung cancer. The model integrates advanced feature extraction and optimization techniques, indicating significant potential for early cancer detection. The proposed Q-GBGWO-ELM model contributes to a more innovative diagnostic approach in clinical practice by offering enhanced precision, efficiency, and adaptability across multiple cancer types. This advancement supports a shift toward more personalized and rapid diagnostic procedures, aiming to improve patient outcomes and reshape current cancer care practices with AI-driven accuracy and efficiency.

Meteorological Visibility Estimation Using Landmark Object Extraction and the ANN Method

Visibility can be interpreted as the largest distance of an object that can be recognized or detected under a bright environment that can be used as an environmental indicator for weather conditions and air pollution. The accuracy of the classical approach of visibility calculation, in which meteorological laws and image feature extraction from digital images are used, depends on the quality and noise disturbances of the image. Therefore, artificial intelligence (AI) and digital image approaches have been proposed for visibility estimation in the past. Image features for the whole digital image are generated by pre-trained convolutional neural networks, and the Artificial Neural Network (ANN) is designed for correlation between image features and visibilities. Instead of using the information of the whole digital images, past research has been proposed to identify effective subregions from which image features are generated. A generalized regression neural network (GRNN) was designed to correlate the image features with the visibilities. Past research results showed that this method is more accurate than the classical approach of using handcrafted features. However, the selection of effective subregions of digital images is not fully automated and is based on manual selection by expert judgments. In this paper, we proposed an automatic effective subregion selection method using landmark object extraction techniques. Image features are generated from these LMO subregions, and the ANN is designed to approximate the mapping between LMO regions’ feature values and visibility values. The experimental results show that this approach can minimize the reductant information for ANN training and improve the accuracy of visibility estimation as compared to the single image approach.

Integration of CNN Models and Machine Learning Methods in Credit Score Classification: 2D Image Transformation and Feature Extraction

Abstract The problem of accurately classifying credit scores is critical for financial institutions to assess individual creditworthiness and effectively manage credit risk. Traditional methods often face limitations when processing large datasets, resulting in lower accuracy and longer processing time. To address this issue, this paper proposes a novel approach to credit score classification by integrating convolutional neural networks (CNN) with machine learning methods. First, a 1D dataset of sequential text data is transformed into 2D greyscale images to use 2D CNN models for feature extraction and classification. Six CNN architectures—DenseNet201, GoogLeNet, MobileNetV2, ResNet18, ShuffleNet, and SqueezeNet—are implemented, and the features in the last layer (1000 features) of each CNN are classified using the softmax method. To further improve the performance, the two best CNN models were selected, and a new fully connected layer (NewFC) was added. A class-based feature set [3 × 31,695] representing three credit score types (good, poor, and standard) was extracted from each model and merged into a feature set [6 × 31,695]. This combined feature set was then reclassified using KNN, LDA, Naive Bayes, and SVM algorithms. The performance of both CNN and machine learning methods was evaluated using accuracy, precision, sensitivity, specificity, and F-score metrics. To optimize classification performance and reduce computational cost, the RelieF algorithm was used to select the best 5 out of 6 features. Compared to using all 6 features, significant improvements in accuracy and efficiency were observed, demonstrating the effectiveness of the proposed method in credit score classification.

Personalizing Multimedia Content Recommendations for Intelligent Vehicles Through Text–Image Embedding Approaches

The ability to automate and personalize the recommendation of multimedia contents to consumers has been gaining significant attention recently. The burgeoning demand for digitization and automation of formerly analog communication processes has caught the attention of researchers and professionals alike. In light of the recent interest and anticipated transition to fully autonomous vehicles, this study proposes a text–image embedding method recommender system for the optimization of personalized multimedia content for in-vehicle infotainment. This study leverages existing pre-trained text embedding models and pre-trained image feature extraction methods. Previous research to date has focused mainly on textual-only or image-only analyses. By employing similarity measurements, this study demonstrates how recommendation of the most relevant multimedia content to consumers is enhanced through text–image embedding.

FG-UNet: fine-grained feature-guided UNet for segmentation of weeds and crops in UAV images.

Semantic segmentation of weed and crop images is a key component and prerequisite for automated weed management. For weeds in unmanned aerial vehicle (UAV) images, which are usually characterized by small size and easily confused with crops at early growth stages, existing semantic segmentation models have difficulties to extract sufficiently fine features. This leads to their limited performance in weed and crop segmentation of UAV images. We proposed a fine-grained feature-guided UNet, named FG-UNet, for weed and crop segmentation in UAV images. Specifically, there are two branches in FG-UNet, namely the fine-grained feature branch and the UNet branch. In the fine-grained feature branch, a fine feature-aware (FFA) module was designed to mine fine features in order to enhance the model's ability to segment small objects. In the UNet branch, we used an encoder-decoder structure to realize high-level semantic feature extraction in images. In addition, a contextual feature fusion (CFF) module was designed for the fusion of the fine features and high-level semantic features, thus enhancing the feature discrimination capability of the model. The experimental results showed that our proposed FG-UNet, achieved state-of-the-art performance compared to other semantic segmentation models, with mean intersection over union (MIOU) and mean pixel accuracy (MPA) of 88.06% and 92.37%, respectively. The proposed method in this study lays a solid foundation for accurate detection and intelligent management of weeds. It will have a positive impact on the development of smart agriculture. © 2024 Society of Chemical Industry.

An Enhanced Convolutional Neural Network (CNN) based P-EDR Mechanism for Diagnosis of Diabetic Retinopathy (DR) using Machine Learning

This study focuses on Diabetic Retinopathy (DR), a disease caused by diabetes that affects the retina of the eye and eventually leads to blindness. Diabetes development progresses to retinopathy and must be addressed at an early stage for effective treatment. Currently, DR is classified as Non-Proliferative DR (NPDR) and Proliferative DR (PDR). This study proposes an Enhanced DR (P-EDR) method based on CNN using a high-resolution dataset benchmark of retinal images. Initially, the data were preprocessed by normalization, augmentation, and resizing to improve image quality and feature extraction. Evaluation was based on accuracy, specificity, sensitivity, and AUC-ROC. The proposed CNN-based P-EDR outperformed advanced ML strategies such as Support Vector Machine (SVM), Random Forest (RF), Probabilistic Neural network (PNN), and Gradient Boosting Machine (GBM) that were executed and compared to diagnose and classify DR. The proposed P-EDR extracts features such as a hemorrhage of the NPDR retina image to identify the disease using image processing for classification. P-EDR provides significant features from images in detection and classification, making it a successful model for diagnosing DR with improved accuracy of 93%, sensitivity of 92%, specificity of 94%, and AUC-ROC of 0.97%. These results highlight the potential of a P-EDR-based machine learning model to support ophthalmologists with the early and precise detection of DR, eventually helping with appropriate treatment and prevention of vision loss.

STRAT: Image style transfer with region-aware transformer

STRAT: Image style transfer with region-aware transformer

Feature extraction and fusion algorithm for infrared visible light images based on residual and generative adversarial network

Feature extraction and fusion algorithm for infrared visible light images based on residual and generative adversarial network

Integration of AI Functions for Automatic Identification of Morphological Parameters in Medical Images Using MATLAB

Introduction: Artificial intelligence (AI) is transforming medical imaging by automating complicated and challenging processes with remarkable accuracy. This paper examines the adoption of AI algorithms using MATLAB for the automated recognition of morphological structures in medical imaging. Methods: This research utilizes advanced machine learning (ML) and deep learning (DL) methods to address essential tasks like image segmentation, feature extraction, and parameter estimation. MATLAB's intuitive interface and powerful computational capabilities make it easy to create and test models for specific medical imaging tasks. Segmentation and parameter prediction are carried out with great accuracy in experiments carried out on publically available datasets, such as brain MRI and lung CT scans. In many instances, the Dice coefficients surpass 0.9. While tackling issues like data quality, model interpretability, and computational efficiency, this study emphasizes the benefits of MATLAB for developing and implementing AI solutions in healthcare. Results: The results emphasize MATLAB's capability to improve diagnostic accuracy and operational effectiveness in clinical environments, facilitating future advancements in customized medicine. There has to be effective software for automated analysis and interpretation of the ever-increasing amount of medical imaging data. Conclusions: The purpose of this research is to examine the feasibility of using MATLAB's AI features for the automated detection of morphological parameters in medical images. Here, we compare AI-driven algorithms to more conventional image processing approaches for typical morphological metrics like area, perimeter, shape descriptors, and object recognition, and we find that the former are more accurate and efficient. To verify the efficacy of the suggested AI model, statistical analysis is carried out on a dataset consisting of medical photographs.

Ultra-fast computation of fractal dimension for RGB images

The fractal dimension (FD) is a quantitative parameter widely used to analyze digital images in many application fields such as image segmentation, feature extraction, object recognition, texture analysis, and image compression and denoising, among many others. A variety of algorithms have been previously proposed for estimating the FD, however most of them are limited to binary or gray-scale images only. In recent years, several authors have proposed algorithms for computing the FD of color images. Nevertheless, almost all these methods are computationally inefficient when analyzing large images. Nowadays, color images can be very large in size, and there is a growing trend toward even larger datasets. This implies that the time required to calculate the FD of such datasets can become extremely long. In this paper we present a very efficient GPU algorithm, implemented in CUDA, for computing the FD of RGB color images. Our solution is an extension to RGB of the differential box-counting (DBC) algorithm for gray-scale images. Our implementation simplifies the box-counting computation to very simple operations which are easily combined across iterations. We evaluated our algorithm on two distinct hardware/software platforms using a set of images of increasing size. The performance of our method was compared against two recent FD algorithms for RGB images: a fast box-merging GPU algorithm, and the most advanced approach based on extending the DBC method. The results showed that our GPU algorithm performed very well and achieved speedups of up to 7.9× and 6172.6× regarding these algorithms, respectively. In addition, our algorithm achieved average error rates similar to those obtained by the two reference algorithms when estimating the FD for synthetic images with known FD values, and even outperformed them when processing large images. These results suggest that our GPU algorithm offers a highly reliable and ultra-fast solution for estimating the FD of color images.

Remote Sensing Cross-Modal Text-Image Retrieval Based on Attention Correction and Filtering

Remote sensing cross-modal text-image retrieval constitutes a pivotal component of multi-modal retrieval in remote sensing, central to which is the process of learning integrated visual and textual representations. Prior research predominantly emphasized the overarching characteristics of remote sensing images, or employed attention mechanisms for meticulous alignment. However, these investigations, to some degree, overlooked the intricacies inherent in the textual descriptions accompanying remote sensing images. In this paper, we introduce a novel cross-modal retrieval model, specifically tailored for remote sensing image-text, leveraging attention correction and filtering mechanisms. The proposed model is architected around four primary components: an image feature extraction module, a text feature extraction module, an attention correction module, and an attention filtering module. Within the image feature extraction module, the Visual Graph Neural Network (VIG) serves as the principal encoder, augmented by a multi-tiered node feature fusion mechanism. This ensures a comprehensive understanding of remote sensing images. For text feature extraction, both the Bidirectional Gated Recurrent Unit (BGRU) and the Graph Attention Network (GAT) are employed as encoders, furnishing the model with an enriched understanding of the associated text. The attention correction segment minimizes potential misalignments in image-text pairings, specifically by modulating attention weightings in cases where there’s a unique correlation between visual area attributes and textual descriptors. Concurrently, the attention filtering segment diminishes the influence of extraneous visual sectors and terms in the image-text matching process, thereby enhancing the precision of cross-modal retrieval. Extensive experimentation carried out on both the RSICD and RSITMD datasets, yielded commendable results, attesting to the superior efficacy of the proposed methodology in the domain of remote sensing cross-modal text-image retrieval.

Self-critical strategy adjustment based artificial intelligence method in generating diagnostic reports of respiratory diseases

Objective. Humanity faces many health challenges, among which respiratory diseases are one of the leading causes of human death. Existing AI-driven pre-diagnosis approaches can enhance the efficiency of diagnosis but still face challenges. For example, single-modal data suffer from information redundancy or loss, difficulty in learning relationships between features, and revealing the obscure characteristics of complex diseases. Therefore, it is critical to explore a method that can assist clinicians in detecting lesions early and in pre-diagnosing corresponding diseases.Approach.This paper introduces a novel network structure, strong constraint self-critical strategy network (SCSCS-Net), which can effectively extract image features from chest x-ray images and generate medical image descriptions, assist clinicians in analyzing patients' medical imaging information, deeply explore potential disease characteristics, and assist in making pre-diagnostic decisions. The SCSCS-Net consists of a reinforced cross-modal feature representation model and a self-critical cross-modal alignment model, which are responsible for learning the features interdependence between images and reports by using a multi-subspace self-attention structure and guiding the model in learning report generation strategies to improve the professionalism and consistency of medical terms in generated reports, respectively.Main results.We further compare our model with some advanced models on the same dataset, and the results demonstrate that our method achieves better performance. Finally, the CE and NLG metrics further confirm that the proposed method acquires the ability to generate high-quality medical reports with higher clinical consistency in generating medical reports.Significance.Our novel method has the potential to improve the early detection and pre-diagnosis of respiratory diseases. The model proposed in this paper allows to narrow the gap between artificial intelligence technology and clinical medical diagnosis and provides the possibility for in-depth integration.

Supervised Face Tampering Detection Based on Spatial Channel Attention Mechanism

Face images hold exceptional significance in contemporary society, serving as direct identifiers due to their rich personal attributes, enhancing daily life and work efficiency. However, advancements in deep learning and image processing have led to the proliferation of sophisticated face forgery software, rendering detection increasingly challenging. We propose a novel face tampering detection method utilizing a spatial attention-enhanced bidirectional convolutional neural network to address this. This approach synergizes the strengths of dense convolutional and depthwise separable networks for superior image feature extraction, thereby improving the accuracy of authentic and manipulated face detection. Furthermore, the network is trained to initially localize tampered regions within face images by integrating a spatial channel-based attention module as supervisory input. On three widely used public face forgery datasets, our method achieves an AUC of no less than 96.45%. The experimental results validate the effectiveness of our method in accurately detecting and initially localizing face tampering.

Supervised Functional Principal Component Analysis Under the Mixture Cure Rate Model: An Application to Alzheimer'S Disease.

Brain imaging data is one of the primary predictors for assessing the risk of Alzheimer's disease (AD). This study aims to extract image-based features associated with the possibly right-censored time-to-event outcomes and to improve predictive performance. While the functional proportional hazards model is well-studied in the literature, these studies often do not consider the existence of patients who have a very low risk and are approximately insusceptible to AD. We introduce a functional mixture cure rate model that extends the proportional hazards model by allowing a proportion of event-free patients. We propose a novel supervised functional principal component analysis (sFPCA) method to extract image features associated with AD risk while accounting for the complexity arising from right censoring. The proposed method accommodates the irregular boundary issue inherent in brain images with bivariate splines over triangulations. We demonstrate the advantages of the proposed method through extensive simulation studies and provide an application to the Alzheimer's Disease Neuroimaging Initiative (ADNI) study.

Recognition of Foal Nursing Behavior Based on an Improved RT-DETR Model

Foal nursing behavior is a crucial indicator of healthy growth. The mare being in a standing posture and the foal being in a suckling posture are important markers for foal suckling behavior. To enable the recognition of a mare’s standing posture and its foal’s suckling posture in stalls, this paper proposes an RT-DETR-Foalnursing model based on RT-DETR. The model employs SACGNet as the backbone to enhance the efficiency of image feature extraction. Furthermore, by incorporating a multiscale multihead attention module and a channel attention module into the Adaptive Instance Feature Integration (AIFI), the model strengthens feature utilization and integration capabilities, thereby improving recognition accuracy. Experimental results demonstrate that the improved RT-DETR achieves a best mAP@50 of 98.5%, increasing by 1.8% compared to the RT-DETR. Additionally, this study achieves real-time statistical analysis of the duration of the foal in the suckling posture, which is one of the important indicators for determining whether the foal is suckling. This has significant implications for the healthy growth of foals.

MM-PCQA+: Advancing Multi-Modal Learning for Point Cloud Quality Assessment

The importance of visual quality in point clouds has been significantly underlined due to the rapid rise in 3D vision applications which aim to deliver affordable and superior user experiences. Reviewing the evolution of point cloud quality assessment (PCQA), it’s observed that visual quality evaluation typically employs single-modal data, either sourced from 2D projections or the 3D point clouds. The 2D projections possess abundant texture and semantic information while they are heavily reliant on viewpoints. In contrast, 3D point clouds are more reactive to geometric distortions and viewpoint-invariant. Consequently, to maximize the benefits of both point cloud and image modalities, we present an advanced no-reference Multi-Modal Point Cloud Quality Assessment (MM-PCQA+) metric. Specifically, we divide the point clouds into sub-models to reflect local geometric distortions such as point shifting and down-sampling. Afterwards, we render the point clouds using a cube-like projection setup and sample the projections of interest using a point-visible-ratio for image feature extraction. In order to fulfill these objectives, the sub-models and projected images are encoded using point-based and image-based neural networks. Lastly, we implement symmetric cross-modal attention to amalgamate multi-modal quality-aware features. Experimental results demonstrate that our metric surpasses all state-of-the-art methods and significantly advances beyond previous no-reference PCQA methods.

Efficient guided inpainting of larger hole missing images based on hierarchical decoding network

When dealing with images containing large hole-missing regions, deep learning-based image inpainting algorithms often face challenges such as local structural distortions and blurriness. In this paper, a novel hierarchical decoding network for image inpainting is proposed. Firstly, the structural priors extracted from the encoding layer are utilized to guide the first decoding layer, while residual blocks are employed to extract deep-level image features. Secondly, multiple hierarchical decoding layers progressively fill in the missing regions from top to bottom, then interlayer features and gradient priors are used to guide information transfer between layers. Furthermore, a proposed Multi-dimensional Efficient Attention is introduced for feature fusion, enabling more effective extraction of image features across different dimensions compared to conventional methods. Finally, Efficient Context Fusion combines the reconstructed feature maps from different decoding layers into the image space, preserving the semantic integrity of the output image. Experiments have been conducted to validate the effectiveness of the proposed method, demonstrating superior performance in both subjective and objective evaluations. When inpainting images with missing regions ranging from 50% to 60%, the proposed method achieves improvements of 0.02 dB (0.22 dB) and 0.001 (0.003) in PSNR and SSIM, on the CelebA-HQ (Places2) dataset, respectively.

BurgsVO: Burgs-Associated Vertex Offset Encoding Scheme for Detecting Rotated Ships in SAR Images

Synthetic Aperture Radar (SAR) is a crucial remote sensing technology with significant advantages. Ship detection in SAR imagery has garnered significant attention. However, existing ship detection methods often overlook feature extraction, and the unique imaging mechanisms of SAR images hinder the direct application of conventional natural image feature extraction techniques. Moreover, oriented bounding box-based detection methods often prioritize accuracy excessively, leading to increased parameters and computational costs, which in turn elevate computational load and model complexity. To address these issues, we propose a novel two-stage detector, Burgs-rooted vertex offset encoding scheme (BurgsVO), for detecting rotated ships in SAR images. BurgsVO consists of two key modules: the Burgs equation heuristics module, which facilitates feature extraction, and the average diagonal vertex offset (ADVO) encoding scheme, which significantly reduces computational costs. Specifically, the Burgs equation module integrates temporal information with spatial data for effective feature aggregation, establishing a strong foundation for subsequent object detection. The ADVO encoding scheme reduces parameters through anchor transformation, leveraging geometric similarities between quadrilaterals and triangles to further reduce computational costs. Experimental results on the RSSDD and RSDD benchmarks demonstrate that the proposed BurgsVO outperforms the state-of-the-art detectors in both accuracy and efficiency.

Research on Ship-Type Recognition Based on Fusion of Ship Trajectory Image and AIS Time Series Data

Achieving accurate and efficient ship-type recognition is crucial for the development and management of modern maritime traffic systems. To overcome the limitations of existing methods that rely solely on AIS time series data or navigation trajectory images as single-modal approaches, this study introduces TrackAISNet, a multimodal ship classification model that seamlessly integrates ship trajectory images with AIS time series data for improved performance. The model employs a parallel structure, utilizing a lightweight neural network to extract features from trajectory images, and a specially designed TCN-GA (Temporal Convolutional Network with Global Attention) to capture the temporal dependencies and long-range relationships in the AIS time series data. The extracted image features and temporal features are then fused, and the combined features are fed into a classification network for final classification. We conducted experiments on a self-constructed dataset of variable-length AIS time series data comprising four types of ships. The results show that the proposed model achieved an accuracy of 81.38%, recall of 81.11%, precision of 80.95%, and an F1 score of 81.38%, outperforming the benchmark single-modal algorithms. Additionally, on a publicly available dataset containing three types of fishing vessel operations, the model demonstrated improvements in accuracy, recall, and F1 scores by 5.5%, 4.88%, and 5.88%, respectively.

Multi-objective classification prediction method for coal gangue based on image feature extraction

ABSTRACT Coal gangue is a solid waste discharged during the coal production process, consisting of rocks and minerals, and is a nonmetal mineral resource with attributes suitable for comprehensive utilization. This study explores the pre-classification issue of coal gangue comprehensive utilization through image recognition methods. Initially, coal gangue was categorized into four types – Residual coal, Gray gangue, Red gangue, and White gangue-based on its appearance, chemical composition, phase composition, and usage characteristics. Subsequently, the color features, texture features, and shape features of different types of coal gangue were analyzed using box plots, correlation coefficients, and cluster analysis. Finally, by comprehensively considering color, texture, and shape features, and by simplifying highly correlated variables, a support vector machine was employed for classification prediction. The overall recognition accuracy rate for multi-target classification reached 89%.