Search Results (6)

The escalating conflict between cetaceans and fisheries underscores the need for efficient mitigation strategies that balance conservation priorities with economic viability. This study presents a TinyML-driven approach deploying an optimized Convolutional Neural Network (CNN) on a Raspberry Pi Zero 2 W for real-time detection of bottlenose dolphin whistles, leveraging spectrogram analysis to address acoustic monitoring challenges. Specifically, a CNN model previously developed for classifying dolphins’ vocalizations and originally implemented with TensorFlow was converted to TensorFlow Lite (TFLite) with architectural optimizations, reducing the model size by 76%. Both TensorFlow and TFLite models were trained on 22 h of underwater recordings taken in controlled environments and processed into 0.8 s spectrogram segments (300 × 150 pixels). Despite reducing model size, TFLite models maintained the same accuracy as the original TensorFlow model (87.8% vs. 87.0%). Throughput and latency were evaluated by varying the thread allocation (1–8 threads), revealing the best performance at 4 threads (quad-core alignment), achieving an inference latency of 120 ms and sustained throughput of 8 spectrograms/second. The system demonstrated robustness in 120 h of continuous stress tests without failure, underscoring its reliability in marine environments. This work achieved a critical balance between computational efficiency and detection fidelity (F1-score: 86.9%) by leveraging quantized, multithreaded inference. These advancements enable low-cost devices for real-time cetacean presence detection, offering transformative potential for bycatch reduction and adaptive deterrence systems. This study bridges artificial intelligence innovation with ecological stewardship, providing a scalable framework for deploying machine learning in resource-constrained settings while addressing urgent conservation challenges. Full article

This paper aims to assist novice gardeners in identifying plant diseases to circumvent misdiagnosing their plants and to increase general horticultural knowledge for better plant growth. In this paper, we develop a mobile plant care support system (“AgroAId”), which incorporates computer vision technology to classify a plant’s [species–disease] combination from an input plant leaf image, recognizing 39 [species-and-disease] classes. Our method comprises a comparative analysis to maximize our multi-label classification model’s performance and determine the effects of varying the convolutional neural network (CNN) architectures, transfer learning approach, and hyperparameter optimizations. We tested four lightweight, mobile-optimized CNNs—MobileNet, MobileNetV2, NasNetMobile, and EfficientNetB0—and tested four transfer learning scenarios (percentage of frozen-vs.-retrained base layers): (1) freezing all convolutional layers; (2) freezing 80% of layers; (3) freezing 50% only; and (4) retraining all layers. A total of 32 model variations are built and assessed using standard metrics (accuracy, F1-score, confusion matrices). The most lightweight, high-accuracy model is concluded to be an EfficientNetB0 model using a fully retrained base network with optimized hyperparameters, achieving 99% accuracy and demonstrating the efficacy of the proposed approach; it is integrated into our plant care support system in a TensorFlow Lite format alongside the front-end mobile application and centralized cloud database. Finally, our system also uses the collective user classification data to generate spatiotemporal analytics about regional and seasonal disease trends, making these analytics accessible to all system users to increase awareness of global agricultural trends. Full article

The importance of an embedded wearable device with automatic detection and alarming cannot be overstated, given that 15–30% of patients with atrial fibrillation are reported to be asymptomatic. These asymptomatic patients do not seek medical care, hence traditional diagnostic tools including Holter are not effective for the further prevention of associated stroke or heart failure. This is likely to be more so in the era of COVID-19, in which patients become more reluctant on hospitalization and checkups. However, little literature is available on this important topic. For this reason, this study developed efficient deep learning with model compression, which is designed to use ECG data and classify arrhythmia in an embedded wearable device. ECG-signal data came from Korea University Anam Hospital in Seoul, Korea, with 28,308 unique patients (15,412 normal and 12,896 arrhythmia). Resnets and Mobilenets with model compression (TensorFlow Lite) were applied and compared for the diagnosis of arrhythmia in an embedded wearable device. The weight size of the compressed model registered a remarkable decrease from 743 MB to 76 KB (1/10000), whereas its performance was almost the same as its original counterpart. Resnet and Mobilenet were similar in terms of accuracy, i.e., Resnet-50 Hz (97.3) vs. Mo-bilenet-50 Hz (97.2), Resnet-100 Hz (98.2) vs. Mobilenet-100 Hz (97.9). Here, 50 Hz/100 Hz denotes the down-sampling rate. However, Resnets took more flash memory and longer inference time than did Mobilenets. In conclusion, Mobilenet would be a more efficient model than Resnet to classify arrhythmia in an embedded wearable device. Full article

Background and Objectives—Coexisting tree and shrub species will have to withstand more arid conditions as temperatures keep rising in the Mediterranean Basin. However, we still lack reliable assessments on how climate and drought affect the radial growth of tree and shrub species at intra- and interannual time scales under semi-arid Mediterranean conditions. Materials and Methods—We investigated the growth responses to climate of four co-occurring gymnosperms inhabiting semi-arid Mediterranean sites in northeastern Spain: two tree species (Aleppo pine, Pinus halepensis Mill.; Spanish juniper, Juniperus thurifera L.) and two shrubs (Phoenicean juniper, Juniperus phoenicea L.; Ephedra nebrodensis Tineo ex Guss.). First, we quantified the intra-annual radial-growth rates of the four species by periodically sampling wood samples during one growing season. Second, we quantified the climate–growth relationships at an interannual scale at two sites with different soil water availability by using dendrochronology. Third, we simulated growth responses to temperature and soil moisture using the forward, process-based Vaganov‒Shashkin (VS-Lite) growth model to disentangle the main climatic drivers of growth. Results—The growth of all species peaked in spring to early summer (May–June). The pine and junipers grew after the dry summer, i.e., they showed a bimodal growth pattern. Prior wet winter conditions leading to high soil moisture before cambium reactivation in spring enhanced the growth of P. halepensis at dry sites, whereas the growth of both junipers and Ephedra depended more on high spring–summer soil moisture. The VS-Lite model identified these different influences of soil moisture on growth in tree and shrub species. Conclusions—Our approach (i) revealed contrasting growth dynamics of co-existing tree and shrub species under semi-arid Mediterranean conditions and (ii) provided novel insights on different responses as a function of growth habits in similar drought-prone regions. Full article

There is a lack of knowledge on how tree species respond to climatic constraints like water shortages and related atmospheric patterns across broad spatial and temporal scales. These assessments are needed to project which populations will better tolerate or respond to global warming across the tree species distribution range. Warmer and drier conditions have been forecasted for the Mediterranean Basin, where Aleppo pine (Pinus halepensis Mill.) is the most widely distributed conifer in dry sites. This species shows plastic growth responses to climate, being particularly sensitive to drought. We evaluated how 32 Aleppo pine forests responded to climate during the second half of the 20th century by using dendrochronology. Climatic constraints of radial growth were inferred by fitting the Vaganov–Shashkin (VS-Lite) growth model to ring-width data from our Aleppo pine forest network. Our findings reported that Aleppo pine growth decreased and showed the highest common coherence among trees in dry, continental sites located in southeastern and eastern inland Spain and Algeria. In contrast, growth increased in wetter sites located in northeastern Spain. Overall, across the Aleppo pine network tree growth was enhanced by prior wet winters and cool and wet springs, whilst warm summers were associated with less growth. The relationships between site ring-width chronologies were higher in nearby forests. This explains why Aleppo pine growth was distinctly linked to indices of atmospheric circulation patterns depending on the geographical location of the forests. The western forests were more influenced by moisture and temperature conditions driven by the Western Mediterranean Oscillation (WeMO) and the Northern Atlantic Oscillation (NAO), the southern forests by the East Atlantic (EA) and the august NAO, while the Balearic, Tunisian and northeastern sites by the Arctic Oscillation (AO) and the Scandinavian pattern (SCA). The climatic constraints for Aleppo pine tree growth and its biogeographical variability were well captured by the VS-Lite model. The model performed better in dry and continental sites, showing strong growth coherence between trees and climatic limitations of growth. Further research using similar broad-scale approaches to climate–growth relationships in drought-prone regions deserves more attention. Full article

Drought stress has induced dieback episodes affecting many forest types and tree species worldwide. However, there is scarce information regarding drought-triggered growth decline and canopy dieback in Mediterranean deciduous oaks. These species face summer drought but have to form new foliage every spring which can make them vulnerable to hotter and drier conditions during that season. Here, we investigated two stands dominated by Quercus frainetto Ten. and Quercus canariensis Willd. and situated in southern Italy and Spain, respectively, showing drought-induced dieback since the 2000s. We analyzed how radial growth and its responses to climate differed between non-declining (ND) and declining (D) trees, showing different crown defoliation and coexisting in each stand by: (i) characterizing growth variability and its responsiveness to climate and drought through time, and (ii) simulating growth responses to soil moisture and temperature thresholds using the Vaganov–Shashkin VS-lite model. Our results show how growth responsiveness to climate and drought was higher in D trees for both oak species. Growth has become increasingly limited by warmer-drier climate and decreasing soil moisture availability since the 1990s. These conditions preceded growth drops in D trees indicating they were more vulnerable to warming and aridification trends. Extremely warm and dry conditions during the early growing season trigger dieback. Changes in the seasonal timing of water limitations caused contrasting effects on long-term growth trends of D trees after the 1980s in Q. frainetto and during the 1990s in Q. canariensis. Using growth models allows identifying early-warning signals of vulnerability, which can be compared with shifts in the growth responses to warmer and drier conditions. Our approach facilitates establishing drought-vulnerability thresholds by combining growth models with field records of dieback. Full article