Cocoa hybridisation generates new varieties which are resistant to several plant diseases, but has individual chemical characteristics that affect chocolate production. Image analysis is a useful method for visual discrimination of cocoa beans, while deep learning (DL) has emerged as thede factotechnique for image processing . However, these algorithms require a large amount of data and careful tuning of hyperparameters. Since it is necessary to acquire a large number of images to encompass the wide range of agricultural products, in this paper, we compare a Deep Computer Vision System (DCVS) and a traditional Computer Vision System (CVS) to classify cocoa beans into different varieties. For DCVS, we used a Resnet18 and Resnet50 as backbone, while for CVS, we experimented traditional machine learning algorithms, Support Vector Machine (SVM), and Random Forest (RF). All the algorithms were selected since they provide good classification performance and their potential application for food classification A dataset with 1,239 samples was used to evaluate both systems. The best accuracy was 96.82% for DCVS (ResNet 18), compared to 85.71% obtained by the CVS using SVM. The essential handcrafted features were reported and discussed regarding their influence on cocoa bean classification. Class Activation Maps was applied to DCVS’s predictions, providing a meaningful visualisation of the most important regions of the images in the model.

Imaging sensors are largely employed in the food processing industry for quality control. Flour from malting barley varieties is a valuable ingredient in the food industry, but its use is restricted due to quality aspects such as color variations and the presence of husk fragments. On the other hand, naked varieties present superior quality with better visual appearance and nutritional composition for human consumption. Computer Vision Systems (CVS) can provide an automatic and precise classification of samples, but identification of grain and flour characteristics require more specialized methods. In this paper, we propose CVS combined with the Spatial Pyramid Partition ensemble (SPPe) technique to distinguish between naked and malting types of twenty-two flour varieties using image features and machine learning. SPPe leverages the analysis of patterns from different spatial regions, providing more reliable classification. Support Vector Machine (SVM), k-Nearest Neighbors (k-NN), J48 decision tree, and Random Forest (RF) were compared for samples’ classification. Machine learning algorithms embedded in the CVS were induced based on 55 image features. The results ranged from 75.00% (k-NN) to 100.00% (J48) accuracy, showing that sample assessment by CVS with SPPe was highly accurate, representing a potential technique for automatic barley flour classification.

The stream mining paradigm has become increasingly popular due to the vast number of algorithms and methodologies it provides to address the current challenges of Internet of Things (IoT) and modern machine learning systems. Change detection algorithms, which focus on identifying drifts in the data distribution during the operation of a machine learning solution, are a crucial aspect of this paradigm. However, selecting the best change detection method for different types of concept drift can be challenging. This work aimed to provide a benchmark for four drift detection algorithms (EDDM, DDM, HDDMW, and HDDMA) for abrupt, gradual, and incremental drift types. To shed light on the capacity and possible trade-offs involved in selecting a concept drift algorithm, we compare their detection capability, detection time, and detection delay. The experiments were carried out using synthetic datasets, where various attributes, such as stream size, the amount of drifts, and drift duration can be controlled and manipulated on our generator of synthetic stream. Our results show that HDDMW provides the best trade-off among all performance indicators, demonstrating superior consistency in detecting abrupt drifts, but has suboptimal time consumption and a limited ability to detect incremental drifts. However, it outperforms other algorithms in detection delay for both abrupt and gradual drifts with an efficient detection performance and detection time performance.

With the increasing demand for time-series analysis, driven by the proliferation of IoT devices and real-time data-driven systems, detecting change points in time series has become critical for accurate short-term prediction. The variability in patterns necessitates frequent analysis to sustain high performance by acquiring the hyperparameter. The Cumulative Sum (CUSUM) method, based on calculating the cumulative values within a time series, is commonly used for change detection due to its early detection of small drifts, simplicity, low computational cost, and robustness to noise. However, its effectiveness heavily depends on the hyperparameter configuration, as a single setup may not be universally suitable across the entire time series. Consequently, fine-tuning is often required to achieve optimal results, yet this selection process is traditionally performed through trial and error or prior expert knowledge, which introduces subjectivity and inefficiency. To address this challenge, several strategies have been proposed to facilitate hyperparameter optimizations, as traditional methods are impractical. Meta-learning-based techniques present viable alternatives for periodic hyperparameter optimization, enabling the selection of configurations that adapt to dynamic scenarios. This work introduces a meta-modeling scheme designed to automate the recommendation of hyperparameters for the CUSUM algorithm. Benchmark datasets from the literature were used to evaluate the proposed framework. The results indicate that this framework preserves high accuracy while significantly reducing time requirements compared to Grid Search and Genetic Algorithm optimization.

Edge computing (EC) is a promising technology capable of bridging the gap between Cloud computing services and the demands of emerging technologies such as the Internet of Things (IoT). Most EC-based solutions, from wearable devices to smart cities architectures, benefit from Machine Learning (ML) methods to perform various tasks, such as classification. In these cases, ML solutions need to deal efficiently with a huge amount of data, while balancing predictive performance, memory and time costs, and energy consumption. The fact that these data usually come in the form of a continuous and evolving data stream makes the scenario even more challenging. Many algorithms have been proposed to cope with data stream classification, e.g., Very Fast Decision Tree (VFDT) and Strict VFDT (SVFDT). Recently, Online Local Boosting (OLBoost) has also been introduced to improve predictive performance without modifying the underlying structure of the decision tree produced by these algorithms. In this work, we compared the four-way relationship among time efficiency, energy consumption, predictive performance, and memory costs, tuning the hyperparameters of VFDT and the two versions of SVFDT with and without OLBoost. Experiments over 6 benchmark datasets using an EC device revealed that VFDT and SVFDT-I were the most energy-friendly algorithms, with SVFDT-I also significantly reducing memory consumption. OLBoost, as expected, improved the predictive performance, but caused a deterioration in memory and energy consumption.