Comprehensive Analysis of Machine and Deep Learning Approaches for Fine-Grained Classification in Optical Remote Sensing Imagery

Deep learning methods have been shown to significantly improve the performance of building extraction from optical remote sensing imagery. However, keeping the morphological characteristics, especially the boundaries, is still a challenge that requires further study. In this paper, we propose a novel fully convolutional network (FCN) for accurately extracting buildings, in which a boundary learning task is embedded to help maintain the boundaries of buildings. Specifically, in the training phase, our framework simultaneously learns the extraction of buildings and boundary detection and only outputs extraction results while testing. In addition, we introduce spatial variation fusion (SVF) to establish an association between the two tasks, thus coupling them and making them share the latent semantics and interact with each other. On the other hand, we utilize separable convolution with a larger kernel to enlarge the receptive fields while reducing the number of model parameters and adopt the convolutional block attention module (CBAM) to boost the network. The proposed framework was extensively evaluated on the WHU Building Dataset and the Inria Aerial Image Labeling Dataset. The experiments demonstrate that our method achieves state-of-the-art performance on building extraction. With the assistance of boundary learning, the boundary maintenance of buildings is ameliorated.

Classification and prediction tasks are common in health research. With the increasing availability of vast health data repositories (e.g. electronic medical record databases) and advances in computing power, traditional statistical approaches are being augmented or replaced with machine learning (ML) approaches to classify and predict health outcomes. ML describes the automated process of identifying ("learning") patterns in data to perform tasks. Developing an ML model includes selecting between many ML models (e.g. decision trees, support vector machines, neural networks); model specifications such as hyperparameter tuning; and evaluation of model performance. This process is conducted repeatedly to find the model and corresponding specifications that optimize some measure of model performance. ML models can make more accurate classifications and predictions than their statistical counterparts and confer greater flexibility when modelling unstructured data or interactions between covariates; however, many ML models require larger sample sizes to achieve good classification or predictive performance and have been criticized as "black box" for their poor transparency and interpretability. ML holds potential in family medicine for risk profiling of patients' disease risk and clinical decision support to present additional information at times of uncertainty or high demand. In the future, ML approaches are positioned to become commonplace in family medicine. As such, it is important to understand the objectives that can be addressed using ML approaches and the associated techniques and limitations. This article provides a brief introduction into the use of ML approaches for classification and prediction tasks in family medicine.

A machine learning approach for hierarchical classification of software requirements

Intrusion detection is a major challenge for security experts in the cyber world. Traditional IDS failed to detect complex and unknown cyber-attacks. Machine learning has become a vibrant technology for cybersecurity. There exists several machine learning algorithms to detect intrusion. Most classifiers are well suited to detect the attacks. However, improving accuracy and detecting unknown attacks in existing IDSs is a great challenge. Therefore, the detailed comparative study of various machine learning approaches such as artificial neural networks, support vector machine, decision tree, and hybrid classifiers used by researchers for intrusion detection are done. Deep learning is an emerging approach which suits well for large data. Deep learning techniques find optimal feature set and classify low-frequency attacks better than other techniques. This study also summarizes literatures in deep learning approaches such as deep auto-encoder, Boltzmann machine, recurrent neural networks, convolutional neural networks, and deep neural networks. Moreover, the datasets used in various literatures and the analysis of deep learning approaches based on the performance metrics are also done. Future directions to detect intrusion are also provided. This study in fact will be helpful to develop IDS based on artificial intelligence approaches such as machine learning and deep learning.KeywordsMachine learningIntrusion detectionFeature selectionDeep learningCyber securityClassifier

Aircraft detection is the challenging task in optical remote sensing image processing system. Due to the shadow affection, the accuracy of detection results getting lower. Therefore, a method of the land aircraft detection which based on shadow affection in optical remote sensing image has been proposed. First, utilizing the anisotropic diffusion method to filter the noise in image, and analyzing the characteristics of gray-scale image to locate shadow regions and candidate target areas. Then, segmentation is performed on the filter able image to highlight the target. Next, we use the morphology operation to strengthen the target area connectedness. Finally, making full use of the aircraft features for target representation, it will reflecting the aircraft target prior-knowledge, at the same time, the accuracy of detection results will be greatly enhancement. Experimental results reveal the feasibility and validity of the proposed approach in detection aircrafts in optical remote sensing images.

Optimized machine learning approaches for identifying vertical temperature gradient on ballastless track in natural environments

Data Complexity-Oriented Classification of Multispectral Remote Sensing Imagery via Machine and Deep Learning Approaches

Performance optimization of wireless networks is typically complicated because of high computational complexity and dynamic channel conditions. Considering a specific case, the recent introduction of intelligent reflecting surface (IRS) can reshape the wireless channels by controlling the scattering elements' phase shifts, namely, passive beamforming. However, due to the large size of scattering elements, the IRS's beamforming optimization becomes intractable. In this article, we focus on machine learning (ML) approaches for complex optimization problems in wireless networks. ML approaches can provide flexibility and robustness against uncertain and dynamic systems. However, practical challenges still remain due to slow convergence in offline training or online learning. This motivated us to design a novel optimization-driven ML framework that exploits the efficiency of model-based optimization and the robustness of model-free ML approaches. Splitting the control variables into two parts allows one part to be updated by the outer loop ML approach while the other part is solved by the inner loop optimization. The case study in IRS-assisted wireless networks confirms that the optimization-driven ML framework can improve learning efficiency and the reward performance significantly compared to conventional model-free ML approaches.

Human emotions play a key role in numerous decision-making processes. The ability to correctly identify likes and dislikes as well as excitement and boredom would facilitate novel applications in neuromarketing, affective entertainment, virtual rehabilitation and forensic neuroscience that leverage on sub-conscious human affective states. In this neuroinformatics investigation, we seek to recognize human preferences and excitement passively through the use of electroencephalography (EEG) when a subject is presented with some 3D visual stimuli. Our approach employs the use of machine learning in the form of deep neural networks to classify brain signals acquired using a brain-computer interface (BCI). In the first part of our study, we attempt to improve upon our previous work, which has shown that EEG preference classification is possible although accuracy rates remain relatively low at 61%-67% using conventional deep learning neural architectures, where the challenge mainly lies in the accurate classification of unseen data from a cohort-wide sample that introduces inter-subject variability on top of the existing intra-subject variability. Such an approach is significantly more challenging and is known as subject-independent EEG classification as opposed to the more commonly adopted but more time-consuming and less general approach of subject-dependent EEG classification. In this new study, we employ deep networks that allow dropouts to occur in the architecture of the neural network. The results obtained through this simple feature modification achieved a classification accuracy of up to 79%. Therefore, this study has shown that the use of a deep learning classifier was able to achieve an increase in emotion classification accuracy of between 13% and 18% through the simple adoption of the use of dropouts compared to a conventional deep learner for EEG preference classification. In the second part of our study, users are exposed to a roller-coaster experience as the emotional stimuli which are expected to evoke the emotion of excitement, while simultaneously wearing virtual reality goggles, which delivers the virtual reality experience of excitement, and an EEG headset, acquires the raw brain signals detected when exposed to this excitement stimuli. Here, a deep learning approach is used to improve the excitement detection rate to well above the 90% accuracy level. In a prior similar study, the use of conventional machine learning approaches involving k-Nearest Neighbour (kNN) classifiers and Support Vector Machines (SVM) only achieved prediction accuracy rates of between 65% and 89%. Using a deep learning approach here, rates of 78%-96% were achieved. This demonstrates the superiority of adopting a deep learning approach over other machine learning approaches for detecting human excitement when immersed in an immersive virtual reality environment.

This study investigates the feasibility of using tear sample analysis, based on protein corona formation on gold nanoparticles combined with electrospray ionization mass spectrometry (ESI-MS) and machine learning techniques, as a non-invasive approach for the detection of choroidal melanoma. The aim is to assess whether protein-nanoparticle interactions can support early and reliable identification of this ocular condition. Tear samples were collected using Schirmer strips from six healthy individuals and six patients diagnosed with choroidal melanoma, with subsequent augmentation to 18 samples per group. Gold nanoparticles (AuNPs, ~ 20 nm) were synthesized via citrate reduction and incubated with tear samples to form protein coronas, which were analyzed using ESI-MS. Eight statistical and entropy-based features (mean, variance, skewness, kurtosis, Shannon entropy, approximate entropy, sample entropy, and permutation entropy) were extracted from spectral data. Additionally, Continuous Wavelet Transform (CWT) with Mexican hat wavelet was applied to convert mass spectrometry data into 128 × 128 RGB images for deep learning analysis. Classification was performed using traditional machine learning models (Random Forest, Support Vector Machine, Decision Tree, Deep Neural Network) and transfer learning with pre-trained CNNs (VGG16, ResNet50, Xception), evaluated through 5-fold cross-validation. Significant differences in spectral intensity parameters were observed between healthy individuals and choroidal melanoma patients (p < 0.001), with notably lower Mean_Intensity values in cancer patients (56.41 ± 46.06 vs. 111.02 ± 10.01, Cohen’s d = 1.64). While m/z parameters showed moderate differences that didn’t reach statistical significance (p = 0.082), entropy-based features demonstrated strong discriminative power. Among traditional machine learning models, Random Forest achieved the highest accuracy (0.959 ± 0.003) and ROC AUC (0.993 ± 0.000) with remarkable computational efficiency (3.90 s per fold). For deep learning approaches using CWT-generated images, VGG16 demonstrated superior performance (Accuracy: 0.976 ± 0.008, ROC AUC: 0.997 ± 0.002) despite requiring significantly higher computational resources (1349.52 s per fold). This study demonstrates that tear sample analysis using protein corona formation on gold nanoparticles with ESI-MS and advanced machine learning techniques offers a promising non-invasive approach for choroidal melanoma detection with performance metrics that compare favorably to existing methods. The significant differences in spectral intensity parameters between groups suggest distinctive proteomic signatures that can be leveraged for diagnostic purposes. While both traditional machine learning and deep learning approaches achieved exceptional performance, each offers distinct advantages in terms of computational efficiency and feature extraction capabilities.

Efficient fertilizer application is vital for enhancing maize production and profitability in Sub-Saharan Africa, where soil fertility varies widely across regions. This study aimed to develop a machine learning approach for generating site-specific fertilizer recommendations for maize production in Ghana and to evaluate its performance against conventional and semi-mechanistic approaches. A random forest machine learning model was trained on 482 maize yield experiments, consisting of 3136 yield observations collected from 1991 to 2020, to predict maize yield response to different fertilizer rates. The model incorporated multiple explanatory variables, including soil properties, climate conditions, and management practices, to generate fertilizer response curves from which fertilizer recommendations were derived for 14 sites across three agro-ecological zones in Ghana where field validation experiments were conducted. On these sites, the recommendations were compared with recommendations derived from the Quantitative Evaluation of the Fertility of Tropical Soils (QUEFTS), Conventional Fertilizer Dose Response (CFDR), and Updated Conventional Fertilizer Dose Response (UCFDR) approaches and validated through field experiments. The machine learning approach generally recommended lower rates of phosphorus and potassium than the other approaches, while nitrogen recommendations were comparable. In the Guinea Savanna zone, the recommendations from the machine learning approach outperformed those from the other approaches, producing higher mean yields for three out of the four sites in the zone. In the Forest-Savanna Transition (FST) zone, the machine learning model recommendations led to higher mean yields at four sites, while the approaches based on QUEFTS and UCFDR performed best at two other sites. In the Semi-deciduous Forest zone, the recommendations of the QUEFTS approach resulted in the highest mean yields at three sites, and CFDR at one site. Despite high input prices during the period of experimentation, the machine learning approach-based recommendations demonstrated higher net profit margins in the FST zone, suggesting cost-effectiveness in this zone. These findings indicate that site-specific fertilizer recommendations are more efficient than blanket recommendations and that machine learning approaches offer a promising and innovative approach for generating cost-effective, site-specific fertilizer recommendations in tropical climates. • ML and QUEFTS approaches provided site-specific fertilizer recommendations. • ML and QUEFTS outperformed conventional fertilizer recommendation approaches. • CFDR and UCFDR were constrained by generalized, non-site-specific recommendations. • Machine learning supports sustainable fertilizer use.

Chemical heuristics have been fundamental to the advancement of chemistry and materials science. These heuristics are typically established by scientists using knowledge and creativity to extract patterns from limited datasets. Machine learning offers opportunities to perfect this approach using computers and larger datasets. Here, we discuss the relationships between traditional heuristics and machine learning approaches. We show how traditional rules can be challenged by large-scale statistical assessment and how traditional concepts commonly used as features are feeding the machine learning techniques. We stress the waste involved in relearning chemical rules and the challenges in terms of data size requirements for purely data-driven approaches. Our view is that heuristic and machine learning approaches are at their best when they work together.

Passive non-invasive sensing signals from wearable devices and smartphones are typically collected continuously without user input. This passive and continuous data collection makes these signals suitable for moment-by-moment monitoring of health-related outcomes, disease diagnosis, and prediction modeling. A growing number of studies have utilized machine learning (ML) approaches to predict and analyze health indicators and diseases using passive non-invasive signals collected via wearable devices and smartphones. This systematic review identified peer-reviewed journal articles utilizing ML approaches for digital phenotyping and measuring digital biomarkers to analyze, screen, identify, and/or predict health-related outcomes using passive non-invasive signals collected from wearable devices or smartphones. PubMed, PubMed with Mesh, Web of Science, Scopus, and IEEE Xplore were searched for peer-reviewed journal articles published up to June 2024, identifying 66 papers. We reviewed the study populations used for data collection, data acquisition details, signal types, data preparation steps, ML approaches used, digital phenotypes and digital biomarkers, and health outcomes and diseases predicted using these ML techniques. Our findings highlight the promising potential for objective tracking of health outcomes and diseases using passive non-invasive signals collected from wearable devices and smartphones with ML approaches for characterization and prediction of a range of health outcomes and diseases, such as stress, seizure, fatigue, depression, and Parkinson’s disease. Future studies should focus on improving the quality of collected data, addressing missing data challenges, providing better documentation on study participants, and sharing the source code of the implemented methods and algorithms, along with their datasets and methods, for reproducibility purposes.

The techniques of point cloud classification in aquatic environments have various applications such as landslide hazard mapping, recovery of lost objects, underwater infrastructure inspection, exploration of mineral resources on the seabed, underwater cultural heritage documentation, environmental preservation and conservation purposes. This study combines acoustic (Sonar) and laser-based (Lidar) remote sensing technologies in an aquatic environment with two machine and deep learning approaches to illustrate the techniques to identify submerged objects. Firstly, the relative accuracy of the underwater imaging system, the BlueView BV5000 Mechanical Scanning Sonar, is evaluated at close range. Secondly, the supervised CANUPO and RandLA-Net classification approaches are used to classify submerged sonar point clouds. Common objects of interest, namely tyres and chairs, were selected for classification. Relative accuracy measurement results showed a centimetre-level root mean square error (RMSE) value, with good accuracies recorded when the scanner is positioned close to objects. The best results were achieved when the target objects were placed at a minimum distance of 2 m from the acoustic scanner. Subsequently, the results of point cloud classification were satisfactory for both approaches. An overall accuracy of 79.81% and an F1 score of 79.80% were achieved using the CANUPO classification approach. On the other hand, an 80.72% overall accuracy and an 80.63% F1 score were obtained using a RandLA-Net approach. These analyses provide a reasonable framework for the parameters that can be used when applying these techniques in natural aquatic environments.

We present an evaluation of different deep learning and machine learning approaches for tree health classification in the Black Forest, the Harz Mountains, and the Göttinger Forest on a unique, highly accurate tree-level dataset. The multispectral UAV data were collected from eight forest plots with diverse tree species, mostly conifers. As ground truth data (GTD), nearly 1500 tree polygons with related attribute information on the health status of the trees were used. This data were collected during extensive fieldwork using a mobile application and subsequent individual tree segmentation. Extensive preprocessing included normalization, NDVI calculations, data augmentation to deal with the underrepresented classes, and splitting the data into training, validation, and test sets. We conducted several experiments using a classical machine learning approach (random forests), as well as different convolutional neural networks (CNNs)—ResNet50, ResNet101, VGG16, and Inception-v3—on different datasets and classes to evaluate the potential of these algorithms for tree health classification. Our first experiment was a binary classifier of healthy and damaged trees, which did not consider the degree of damage or tree species. The best results of a 0.99 test accuracy and an F1 score of 0.99 were obtained with ResNet50 on four band composites using the red, green, blue, and infrared bands (RGBI images), while VGG16 had the worst performance, with an F1 score of only 0.78. In a second experiment, we also distinguished between coniferous and deciduous trees. The F1 scores ranged from 0.62 to 0.99, with the highest results obtained using ResNet101 on derived vegetation indices using the red edge band of the camera (NDVIre images). Finally, in a third experiment, we aimed at evaluating the degree of damage: healthy, slightly damaged, and medium or heavily damaged trees. Again, ResNet101 had the best performance, this time on RGBI images with a test accuracy of 0.98 and an average F1 score of 0.97. These results highlight the potential of CNNs to handle high-resolution multispectral UAV data for the early detection of damaged trees when good training data are available.