Sürdürülebilir Kalkınma Amaçları

• 1
• 1

• MakaleAçık Erişim2026Scopus SJRQ1/r/munzur/782
  EBA-Net: A hybrid EfficientNetB3-BiLSTM-Attention model for species-level diatom classification

  Balo Utku, Esen Damla; Kutlu, Banu; Utku, Anil

  Species identification in diatoms is important for assessing water bodies, determining environmental quality, and analyzing biodiversity. Nonetheless, diatom classification has proven difficult due to intra-class variation, interclass similarity, and data imbalance in microscopic images. In this regard, we propose a new deep learning approach for species-level diatom classification. Various deep learning models, including traditional convolutional neural networks or CNNs (ResNet50, DenseNet121), state-of-the-art CNNs (EfficientNetB3, Xception, MobileNetV3), transformer-based architectures (ViT-B16, Swin-Tiny), and our proposed EBA-Net (EfficientNetB3-BiLSTM-Attention Network), were extensively compared under identical experimental settings. The EBA-Net integrates the advantages of convolutional feature extraction, sequence modeling, and an attention mechanism to improve classification accuracy. It first uses the EfficientNetB3 model as the backbone network to obtain higher-level features, followed by a BiLSTM layer to capture dependencies among the extracted features. Then, it applies the attention mechanism to the BiLSTM output to highlight the most informative features. Experimental results showed that the proposed EBA-Net model outperforms other models considered in the study, yielding 95.8% accuracy and 85.1% macro F1-score. In addition, Top-3 (98.9%) and Top-5 (99.6%) accuracies confirmed the efficiency of the proposed model for classifying objects from visually similar classes. An analysis of the model's interpretability was conducted using Gradient-weighted Class Activation Mapping, showing that the proposed model's attention is concentrated on relevant morphological features of diatoms. To the best of our knowledge, the suggested approach achieves state-of-the-art accuracy on the data set of diatoms used.

  İncele
• MakaleAçık Erişim2025/r/munzur/6194
  Artificial Intelligence Applications in the Aquaculture Industry: Sustainability, Efficiency and Innovative Solutions

  Kutlu, Banu; Utku, Esen Damla Balo

  Artificial Intelligence (AI) is becoming increasingly important to increase efficiency and sustainability by organizing processes and providing significant improvements in different sectors. AI is reshaping the aquaculture sector. AI is being used to help solve optimization problems in aquaculture, reduce costs, and find sustainable solutions. These scenarios guided the evaluation of the future role of AI in aquaculture and the benefits that AI brings to the aquaculture sector in this study. AI is improving the way fish stocks are monitored, live stocks are assessed, diseases among plants and animals are recognized, water quality is tested, and production lines are controlled. Aquaculture uses AI technologies to help detect diseases early, monitor environmental conditions affecting production, and improve both efficiency and sustainability. In addition, AI is examining and providing solutions to environmental issues such as global warming and water resource management.

  • 
  • 
  AtıfOpenAlex5

  İncele
• MakaleAçık Erişim2025/r/munzur/6146
  Enhancing water quality prediction with artificial intelligence: A hybrid convlstm model for spatio-temporal analysis

  Utku, Anıl; Kutlu, Banu; Utku, Esen Damla Balo

  Water quality plays a vital role in public health, environmental sustainability and ecosystem balance. However, industrialization, urbanization, and agricultural activities cause water pollution to increase, threatening both human health and aquatic ecosystems. Traditional water quality monitoring methods are usually time-consuming, costly, and require manual intervention. Therefore, developing automatic, data-driven, and high-accuracy prediction models is crucial for sustainable water management. This study created a hybrid ConvLSTM model to increase prediction accuracy. The created model was comparatively analyzed with RF, SVR, XGBoost, MLP, CNN, and LSTM. The dataset used includes historical measurement values of chemical pollutants. Input data includes geographical coordinates of sample points, chemical parameter type, and compliance with regulatory standards. Experimental results show that ConvLSTM provides the lowest prediction errors by learning spatial and temporal dependencies and reaches the highest accuracy rate with 0.994 R2 compared to other models.

  İncele