Automated Marine Debris Detection: A Deep Learning and Ensemble Approach with Web Deployment

Computer-aided diagnosis (CADx) of polyps is essential for advancing computed tomography colonography (CTC) with diagnostic capability. In this paper, we present a study of investigating the performance between deep learning and Random Forest (RF) classifier for polyp differentiation in CTC. First, we conducted feature extraction via an extended Haralick model (eHM) to build a total of 30 texture features. The gray level co-occurrence matrix (GLCM) is generated to encode 3D CT image information into a 2D matrix as input to the convolutional neural network (CNN). Then, we split the polyp classification into two state-of-the-art frameworks: the eHM texture features/RF and the GLCM texture matrices/CNN. We evaluated their performances by the merit of area under the curve of receiver operating characteristic using 1,278 polyps (confirmed by pathology). Results demonstrated that by balancing the data, both CNN model and RF classifier can learn or analyze features effectively, and achieve high performance. RF classifier in general outperformed CNN model with a gain of 6.4% (balanced datasets) and 5.4% (unbalanced datasets), showing its effective in feature extraction and analysis for polyp differentiation. However, the performance of CNN got improved through the addition of new data with a gain of 3.6% (balanced datasets) and 3.4% (unbalanced datasets), whereas RF classifier showed no gain when we enlarged datasets. This demonstrated that CNN model have the potential to improve the classification task performance when dealing with larger dataset. This study provided valuable information on how to design experiments to improve CADx of polyps.

Sentiment analysis is one of the active research areas in the field of datamining. Machine learning algorithms are capable to implement sentiment analysis. Due to the capacity of self-learning and massive data handling, most of the researchers are using deep learning neural networks for solving sentiment classification tasks. So, in this paper, a new model is designed under a hybrid framework of machine learning and deep learning which couples Convolutional Neural Network and Random Forest classifier for fine-grained sentiment analysis. The Continuous Bag-of-Word (CBOW) model is used to vectorize the text input. The most important features are extracted by the Convolutional Neural Network (CNN). The extracted features are used by the Random Forest(RF) classifier for sentiment classification. The performance of the proposed hybrid CNNRF model is comparedwith the base model such as Convolutional Neural Network (CNN) and Random Forest (RF) classifier. The experimental result shows that the proposed model far beat the existing base models in terms of classification accuracy and effectively integrated genetically-modified CNN with Random Forest classifier.

Over the past decade, object-based image analysis (OBIA) has gained prominence as a widely adopted method for generating land use/land cover (LULC) maps. This study aims to evaluate the performance of various classification algorithms within the OBIA framework using SPOT-6 satellite imagery. The research methodology involved segmenting the images with the multi-resolution segmentation (MRS) algorithm, followed by the application of convolutional neural networks (CNN), random forest (RF), and support vector machine (SVM) algorithms for classification. The study was conducted in the Perpignan province, located in the Pyrénées-Orientales region of France. After the segmentation stage, CNN, RF, and SVM classifiers were employed to classify the image segments based on both spectral and spatial attributes. The accuracy of the resulting thematic maps was assessed using standard metrics, including overall accuracy (OA), the Kappa coefficient (KC), and the F��score (FS). Of the three classifiers, CNN achieved the highest overall accuracy at 91.28%, outperforming SVM, which attained an OA of 90.50%, and RF, which recorded an OA of 87.28%. Additionally, this study explored the integration of explainable artificial intelligence (AI) techniques, specifically the Shapley Additive Explanations (SHAP) algorithm, to enhance the interpretability of the machine learning models. This approach fosters greater trust, accountability, and acceptance in decision-making processes. By leveraging SHAP values, the study provides deeper insights into the decision-making processes of the CNN, SVM, and RF classifiers, ultimately enhancing the transparency and comprehensibility of these models.

Diabetic retinopathy refers to a state of the human eye that affects retina blood vessels, causing vision impairment or even complete vision loss. In this paper, we classify DR fundus images into the presence and absence of the disease by using a combination of deep learning layers and machine learning algorithms. Machine learning (e.g. random forest and support vector machine (SVM)) and deep learning (e.g. convolutional neural networks (CNNs)) are the most well-known approaches for small and big data, respectively, in image classification tasks. The results are compromised when there is a lack of data. Furthermore, a machine learning algorithm takes less time to train than a deep learning method. As a result, we attempted to develop models that combined machine learning and deep learning approaches. So, three models are proposed in this paper to comparatively study their results. The first model utilises dense layers to classify the data, while the second and third models use SVM and random forest classifiers, respectively. Hence, the model employs CNN and machine learning algorithms to increase the accuracy and efficacy of limited data sets. Usually, it is considered that deep learning algorithms perform better on images, while our results show that random forest outperformed the other approaches. We discovered that combining two learning approaches allows us to get superior outcomes on a short data set without sacrificing accuracy or other measures.KeywordsDiabetic retinopathyDeep learningConvolutional neural networkMachine learning

Comprehensive Study for Breast Cancer Using Deep Learning and Traditional Machine Learning

Amid the worldwide COVID-19 pandemic lockdowns, the closure of educational institutes leads to an unprecedented rise in online learning. For limiting the impact of COVID-19 and obstructing its widespread, educational institutions closed their campuses immediately and academic activities are moved to e-learning platforms. The effectiveness of e-learning is a critical concern for both students and parents, specifically in terms of its suitability to students and teachers and its technical feasibility with respect to different social scenarios. Such concerns must be reviewed from several aspects before e-learning can be adopted at such a larger scale. This study endeavors to investigate the effectiveness of e-learning by analyzing the sentiments of people about e-learning. Due to the rise of social media as an important mode of communication recently, people’s views can be found on platforms such as Twitter, Instagram, Facebook, etc. This study uses a Twitter dataset containing 17,155 tweets about e-learning. Machine learning and deep learning approaches have shown their suitability, capability, and potential for image processing, object detection, and natural language processing tasks and text analysis is no exception. Machine learning approaches have been largely used both for annotation and text and sentiment analysis. Keeping in view the adequacy and efficacy of machine learning models, this study adopts TextBlob, VADER (Valence Aware Dictionary for Sentiment Reasoning), and SentiWordNet to analyze the polarity and subjectivity score of tweets’ text. Furthermore, bearing in mind the fact that machine learning models display high classification accuracy, various machine learning models have been used for sentiment classification. Two feature extraction techniques, TF-IDF (Term Frequency-Inverse Document Frequency) and BoW (Bag of Words) have been used to effectively build and evaluate the models. All the models have been evaluated in terms of various important performance metrics such as accuracy, precision, recall, and F1 score. The results reveal that the random forest and support vector machine classifier achieve the highest accuracy of 0.95 when used with Bow features. Performance comparison is carried out for results of TextBlob, VADER, and SentiWordNet, as well as classification results of machine learning models and deep learning models such as CNN (Convolutional Neural Network), LSTM (Long Short Term Memory), CNN-LSTM, and Bi-LSTM (Bidirectional-LSTM). Additionally, topic modeling is performed to find the problems associated with e-learning which indicates that uncertainty of campus opening date, children’s disabilities to grasp online education, and lagging efficient networks for online education are the top three problems.

Machine learning applications have gained popularity over the years as more advanced algorithms like the deep learning (DL) algorithm are being employed in signal identification, classification and detection of cracks or faults in structures. The DL algorithm has broader applications compared to other machine learning systems and it is a creative algorithm capable of processing data, creating pattern, interpreting information due to its high level of accuracy in pattern recognition under stochastic conditions. This research gives an exposition of DL in diverse areas of operations with a focus on plant weed detection which is inspired by the need to treat a specific class of weed with a particular herbicide. A Convolutional Neural Network (CNN) model was trained through transfer learning on a pre-trained ResNet50 model and the performance was evaluated using a random forest (RF) classifier, the trained model was deployed on a raspberry pi for prediction of the test data. Training accuracies of 99% and 93% were obtained for the CNN and RF classifier respectively. Some recommendations have been proffered to improve inference time such as the use of better embedded systems such as the Nvidia Jetson TX2, synchronizing DL hardware accelerators with appropriate optimization techniques. A prospect of this work would be to incorporate an embedded system, deployed with DL algorithms, on an unmanned aerial vehicle or ground vehicle. Overall, it is revealed from this study that DL is highly efficient in every sector and can improve the accuracy on automatic detection of systems in especially in this era of Industry 4.0.

Forest and land fires occur due to natural or manmade causes and have large impacts. Therefore, rapid burned area mapping is needed to investigate impact losses. Remote sensing satellite imagery is a prominent technology for rapid burned area mapping. However, optical data utilization becomes a challenge due to cloud cover. This study aims to evaluate the burned area model derived from optical data (optical-based classification), synthetic aperture radar (SAR) data (SAR-based classification), and combined SAR and optical data using random forest (RF), multilayer perceptron (MLP), and convolutional neural network (CNN) classifiers. SAR-based change detection parameters, such as radar burn ratio (RBR), radar burn difference (RBD), and gray-level co-occurrence matrix (GLCM) texture features, are used as the input features for RF and MLP classifiers. The results show that the CNN classifier outperforms RF and MLP in the case of optical-based classification and combined optical and SAR data classification, with accuracies of 99.73% and 99.86%, respectively. CNN classifiers are relatively not affected by the contribution of SAR data in cloud-free areas, as they give stable classification results in both classification schemes. Combined with the optical and SAR data classification and SAR-based classification, the contribution of GLCM texture features gives better classification results than using the RBR and RBD features for the areas affected by clouds using the RF and MLP classifiers. The contribution of GLCM texture features also significantly affects the MLP classifier more than the RF classifier in cloud-free areas in both classification schemes.

Mapping irrigated plots is essential for better water resource management. Today, the free and open access Sentinel-1 (S1) and Sentinel-2 (S2) data with high revisit time offers a powerful tool for irrigation mapping at plot scale. Up to date, few studies have used S1 and S2 data to provide approaches for mapping irrigated plots. This study proposes a method to map irrigated plots using S1 SAR (synthetic aperture radar) time series. First, a dense temporal series of S1 backscattering coefficients were obtained at plot scale in VV (Vertical-Vertical) and VH (Vertical-Horizontal) polarizations over a study site located in Catalonia, Spain. In order to remove the ambiguity between rainfall and irrigation events, the S1 signal obtained at plot scale was used conjointly to S1 signal obtained at a grid scale (10 km × 10 km). Later, two mathematical transformations, including the principal component analysis (PCA) and the wavelet transformation (WT), were applied to the several SAR temporal series obtained in both VV and VH polarization. Irrigated areas were then classified using the principal component (PC) dimensions and the WT coefficients in two different random forest (RF) classifiers. Another classification approach using one dimensional convolutional neural network (CNN) was also performed on the obtained S1 temporal series. The results derived from the RF classifiers with S1 data show high overall accuracy using the PC values (90.7%) and the WT coefficients (89.1%). By applying the CNN approach on SAR data, a significant overall accuracy of 94.1% was obtained. The potential of optical images to map irrigated areas by the mean of a normalized differential vegetation index (NDVI) temporal series was also tested in this study in both the RF and the CNN approaches. The overall accuracy obtained using the NDVI in RF classifier reached 89.5% while that in the CNN reached 91.6%. The combined use of optical and radar data slightly enhanced the classification in the RF classifier but did not significantly change the accuracy obtained in the CNN approach using S1 data.

Plant species recognition from images or videos is challenging due to a large diversity of plants, variation in orientation, viewpoint, background clutter, etc. In this paper, plant species recognition is carried out using two approaches, namely, traditional method and deep learning approach. In tr aditional method, feature extraction is carried out using Hu moments (shape features), Haralick texture, local binary pattern (LBP) (texture features) and color channel statistics (color features). The extracted features are classified using different classifiers (linear discriminant analysis, logistic regression, classification and regression tree, naïve Bayes, k-nearest neighbor, random forest and bagging classifier). Also, different deep learning architectures are tested in the context of plant species recognition. Three standard datasets (Folio, Swedish leaf and Flavia) and one real-time dataset (Leaf12) is used. It is observed that, in traditional method, feature vector obtained by the combination of color channel statistics+LBP+Hu+Haralick with Random Forest classifier for Leaf12 dataset resulted in a plant recognition accuracy (rank-1) of 82.38%. VGG 16 Convolutional Neural Network (CNN) architecture with logistic regression resulted in an accuracy of 97.14% for Leaf12 dataset. An accuracy of 96.53%, 96.25% and 99.41% is obtained for Folio, Flavia and Swedish leaf datasets using VGG 19 CNN architecture with logistic regression as a classifier. It is also observed that the VGG (Very large Convolutional Neural Network) CNN models provided a higher accuracy rate compared to traditional methods.

Abstract— Precision farming is being revolutionized by the integration of innovative machine learning and computer vision methods. Identifying and classifying weeds and crops accurately remains a major challenge in this field, which has a direct effect on optimizing the yield as well as sustainability. In this work, an approach to smart weed detection based on deep learning using Convolutional Neural Networks (CNN) for feature learning followed by comparison of classifiers to select the best-performing model is introduced. In our research, InceptionV3 was utilized to extract features, and four classifiers—SoftMax, Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF)— were compared. Among them, the Random Forest classifier performed better than others with a validation accuracy of 99.57% and an F1 score of 0.99. Extending the successful application of crop-weed detection, the model was transferred to a new application: forest fire detection. Employing the same CNN- based feature extraction pipeline and Random Forest classification, our system showed high accuracy on a forest fire dataset. In addition, we implemented a real-time detection system using webcam feeds with a processing speed of around 30 frames per second, making practical deployment for environmental monitoring possible. This study not only confirms the efficacy of the union of CNNs and ensemble learning but also exemplifies the versatility of the model architecture in both agricultural and environmental contexts. Index Terms— Convolutional Neural Networks (CNN), InceptionV3, Random Forest, Weed Detection, Crop Classification, Forest Fire Detection, Real-Time Image Processing, Machine Learning, Deep Learning, Precision Agriculture, Environmental Monitoring, Feature Extraction, Webcam Detection.

Urban areas feature complex and heterogeneous land covers which create challenging issues for tree species classification. The increased availability of high spatial resolution multispectral satellite imagery and LiDAR datasets combined with the recent evolution of deep learning within remote sensing for object detection and scene classification, provide promising opportunities to map individual tree species with greater accuracy and resolution. However, there are knowledge gaps that are related to the contribution of Worldview-3 SWIR bands, very high resolution PAN band and LiDAR data in detailed tree species mapping. Additionally, contemporary deep learning methods are hampered by lack of training samples and difficulties of preparing training data. The objective of this study was to examine the potential of a novel deep learning method, Dense Convolutional Network (DenseNet), to identify dominant individual tree species in a complex urban environment within a fused image of WorldView-2 VNIR, Worldview-3 SWIR and LiDAR datasets. DenseNet results were compared against two popular machine classifiers in remote sensing image analysis, Random Forest (RF) and Support Vector Machine (SVM). Our results demonstrated that: (1) utilizing a data fusion approach beginning with VNIR and adding SWIR, LiDAR, and panchromatic (PAN) bands increased the overall accuracy of the DenseNet classifier from 75.9% to 76.8%, 81.1% and 82.6%, respectively. (2) DenseNet significantly outperformed RF and SVM for the classification of eight dominant tree species with an overall accuracy of 82.6%, compared to 51.8% and 52% for SVM and RF classifiers, respectively. (3) DenseNet maintained superior performance over RF and SVM classifiers under restricted training sample quantities which is a major limiting factor for deep learning techniques. Overall, the study reveals that DenseNet is more effective for urban tree species classification as it outperforms the popular RF and SVM techniques when working with highly complex image scenes regardless of training sample size.

Plants meet machines: Prospects in machine learning for plant biology

In recent years, Electroencephalogram (EEG) based user authentication systems have gained significant interest as an innovative approach for identity verification. EEGs are considered to be a novel biometric attribute due to the individuality of each person’s cerebral activity patterns. This work explores the feasibility and efficiency of utilizing EEG signals, generated in response to emotional stimuli, for user authentication applications, by implementing Machine Learning (ML) and Deep Learning (DL) approaches. Support Vector Machine (SVM), Random Forest (RF) classifier and 1D Convolution Neural Network (CNN) were employed to evaluate and compare the performance of EEG-based user authentication for two publicly available EEG datasets, namely DEAP and DENS database. The performance of EEGbased user authentication was significantly high in LAHV emotional state for DENS dataset, achieving an accuracy of 99.2 % and 92.59 % with SVM and modified 1D CNN, respectively.

The Local Climate Zone (LCZ) classification is already widely used in urban heat island and other climate studies. The current classification method does not incorporate crucial urban auxiliary GIS data on building height and imperviousness that could significantly improve urban-type LCZ classification utility as well as accuracy. This study utilized a hybrid GIS- and remote sensing imagery-based framework to systematically compare and evaluate different machine and deep learning methods. The Convolution Neural Network (CNN) classifier outperforms in terms of accuracy, but it requires multi-pixel input, which reduces the output’s spatial resolution and creates a tradeoff between accuracy and spatial resolution. The Random Forest (RF) classifier performs best among the single-pixel classifiers. This study also shows that incorporating building height dataset improves the accuracy of the high- and mid-rise classes in the RF classifiers, whereas an imperviousness dataset improves the low-rise classes. The single-pass forward permutation test reveals that both auxiliary datasets dominate the classification accuracy in the RF classifier, while near-infrared and thermal infrared are the dominating features in the CNN classifier. These findings show that the conventional LCZ classification framework used in the World Urban Database and Access Portal Tools (WUDAPT) can be improved by adopting building height and imperviousness information. This framework can be easily applied to different cities to generate LCZ maps for urban models.