The widespread adoption of Unmanned Aerial Vehicles (UAVs) or drones has introduced significant security and privacy challenges, particularly concerning unauthorized drone activities near sensitive areas. To address these concerns, we propose a novel drone detection model, WRN-YOLO, which integrates the Wide Residual Network (WRN) architecture with the You Only Look Once (YOLO) object detection framework. This integration enhances feature extraction capabilities, leading to improved detection accuracy. Through comprehensive ablation studies, we have identified the optimal YOLO variant that synergizes with our backbone modifications, ensuring superior performance in diverse scenarios. Recognizing the complexities of real-world environments, we have also developed a synthetic dataset designed to train our WRN-YOLO. This dataset encompasses a variety of challenging conditions, including intricate backgrounds and the presence of confounding elements, to robustly assess the model's efficacy. Experimental results demonstrate that our method significantly outperforms existing models in accurately detecting drones amidst complex scenes, offering a promising solution for real-time UAV threat mitigation. The proposed approach ranked Top 3 in the 8th WOSDETC Drone-vs-Bird Detection Challenge. Our source code and synthetic dataset are publicly available at https://github.com/yjwong1999/IJCNN2025-DvB.

Model-Agnostic Meta-learning (MAML) is a widely adopted few-shot learning (FSL) method designed to mitigate the dependency on large, labeled datasets of deep learning-based methods in medical imaging analysis. However, MAML's reliance on a fixed number of gradient descent (GD) steps for task adaptation results in computational inefficiency and task-level overfitting. To address this issue, we introduce Tra-MAML, which optimizes the balance between model adaptation capacity and computational efficiency through a trapezoidal step scheduler (TRA). The TRA scheduler dynamically adjusts the number of GD steps in the inner optimization loop: initially increasing the steps uniformly to reduce variance, maintaining the maximum number of steps to enhance adaptation capacity, and finally decreasing the steps uniformly to mitigate overfitting. Our evaluation of TraMAML against selected FSL methods across four medical imaging datasets demonstrates its superior performance. Notably, Tra-MAML outperforms MAML by 13.36% on the BreaKHis40X dataset in the 3-way 10-shot scenario.

This project employs artificial intelligence, including machine learning and deep learning, to assess COVID-19 readmission risk in Malaysia. It offers tools to mitigate healthcare resource strain and enhance patient outcomes. This study outlines a methodology for classifying COVID-19 readmissions. It starts with dataset description and pre-processing, while the data balancing was computed through Random Oversampling, Borderline SMOTE, and Adaptive Synthetic Sampling. Nine machine learning and ten deep learning techniques are applied, with five-fold cross-validation for evaluation. Optuna is used for hyperparameter selection, while the consistency in training hyperparameters is maintained. Evaluation metrics encompass accuracy, AUC, and training/inference times. Results were based on stratified five-fold cross-validation and different data-balancing methods. Notably, CatBoost consistently excelled in accuracy and AUC across all tables. Using ROS, CatBoost achieved the highest accuracy (0.9882 ± 0.0020) with an AUC of 1.0000 ± 0.0000. CatBoost maintained its superiority in BSMOTE and ADASYN as well. Deep learning approaches performed well, with SAINT leading in ROS and TabNet leading in BSMOTE and ADASYN. Decision Tree ensembles like Random Forest and XGBoost consistently showed strong performance.

Model-Agnostic Meta-Learning (MAML) is a widely used few-shot learning (FSL) technique that reduces reliance on large, labeled datasets in deep learning for medical imaging analysis. However, MAML requires backpropagating through all feature layers for task adaptation, leading to suboptimal computational efficiency. We propose IMAML-IDCG (ImageNet Model-Agnostic Meta-Learning in Invasive Ductal Carcinoma Grading), which enhances computational efficiency for few-shot grading of Invasive Ductal Carcinoma (IDC) through three key techniques: (1) ImageNet feature reusing, (2) ImageNet partial freezing strategy, and (3) adaptive inner learning rate. IMAML-IDCG is initialized with ImageNet pre-trained weights. During the inner optimization loop, only the model’s classifier head layer is optimized, leveraging prior ImageNet knowledge (ImageNet feature reusing) and employing an adaptive learning rate for improved task adaptation. In the outer optimization loop, IMAML-IDCG selectively fine-tunes the last few model layers to enhance efficiency and reduce overfitting (ImageNet partial freezing strategy). We evaluated IMAML-IDCG using the BreaKHis dataset (7,909 images) as the base dataset, and the BCHI (282 images) and PathoIDCG (3,744 images) datasets as the novel datasets. Our empirical results demonstrate that IMAML-IDCG outperforms MAML and other FSL methods in few-shot IDC grading tasks across various cross-magnification domain settings. Notably, IMAML-IDCG achieves a 14.64% improvement over MAML on the BCHI dataset and a 6.04% improvement on the PathoIDCG 40X dataset when meta-trained with the BreaKHis 40X dataset in the 3-way 5-shot scenario.

Palm oil industry is an important economic resource for Malaysia. However, an oil palm tree disease called Basal Stem Rot has impeded the production of palm oil, which caused significant economic loss at the same time. The oil palm tree disease is caused by a fungus known as Ganoderma Boninense. Infected trees often have little to no symptoms during early stage of infection, which made early detection difficult. Early disease detection is necessary to allow early sanitization and disease control efforts. Using Terrestrial Laser Scanning technology, 88 grey-distribution canopy images of oil palm tree were obtained. The images were pre-processed and augmented before being used for training and testing of the deep learning models. The capabilities of the Convolution Neural Network deep learning models in the classification of dataset into healthy and non-healthy class were tested and the best performing model was identified based on the Macro-F1 score. Fine-tuned DenseNet121 model was the best performing model, recorded a Macro F1-score of 0.798. It was also noted that Baseline model showed a relatively remarkable macro-F1 score of 0.747, which was better than all the feature extractor models and some of the fine-tuned models. However, fine-tuned models suffered from model overfitting due to dataset limitations. For future work, it is recommended to increase the sample size and utilize other CNN architectures to improve the model performance and progress towards detecting Basal Stem Rot at the early stage of infection by classifying sample images into multiple classes.

About one-third of the world's vegetable oil and fat supply is made up of palm oil, of which 75% is consumed as food. Palm oil is a vital economic resource for nations like Malaysia. The Basal Stem Rot disease of oil palm trees is one of many obstacles to the production of palm oil. The infection is brought on by a fungus called Ganoderma Boninense, which colonizes trees. Early detection is difficult since the symptoms of infection are sometimes mild to non-existent. Terrestrial laser scanning was used to collect 88 photos of the oil palm tree's grey-distribution canopy. The photos gathered were pre-processed to enhance the performance of the deep learning model. To train and verify the effectiveness of disease detection, a deep learning model called convolution neural network is used. The performance of disease detection is trained and tested using a convolutional neural network deep learning model, which divides the data into two classes: the healthy class and the non-healthy class. The improved DenseNet121 model reports a Macro F1-score of 0.7983. The model could only separate the images into two classes rather than categorizing the images into distinct infection levels, which is a limitation of our work. In order to investigate the feasibility of early oil palm disease diagnosis, it is advised for future research to undertake multi-class or multi-level classification using deep learning.

Debates persist regarding the impact of Stain Normalization (SN) on recent breast cancer histopathological studies. While some studies propose no influence on classification outcomes, others argue for improvement. This study aims to assess the efficacy of SN in breast cancer histopathological classification, specifically focusing on Invasive Ductal Carcinoma (IDC) grading using Convolutional Neural Networks (CNNs). The null hypothesis asserts that SN has no effect on the accuracy of CNN-based IDC grading, while the alternative hypothesis suggests the contrary. We evaluated six SN techniques, with five templates selected as target images for the conventional SN techniques. We also utilized seven ImageNet pre-trained CNNs for IDC grading. The performance of models trained with and without SN was compared to discern the influence of SN on classification outcomes. The analysis unveiled a p-value of 0.11, indicating no statistically significant difference in Balanced Accuracy Scores between models trained with StainGAN-normalized images, achieving a score of 0.9196 (the best-performing SN technique), and models trained with non-normalized images, which scored 0.9308. As a result, we did not reject the null hypothesis, indicating that we found no evidence to support a significant discrepancy in effectiveness between stain-normalized and non-normalized datasets for IDC grading tasks. This study demonstrates that SN has a limited impact on IDC grading, challenging the assumption of performance enhancement through SN.

Computer-aided Invasive Ductal Carcinoma (IDC) grading classification systems based on deep learning have shown that deep learning may achieve reliable accuracy in IDC grade classification using histopathology images. However, there is a dearth of comprehensive performance comparisons of Convolutional Neural Network (CNN) designs on IDC in the literature. As such, we would like to conduct a comparison analysis of the performance of seven selected CNN models: EfficientNetB0, EfficientNetV2B0, EfficientNetV2B0-21k, ResNetV1-50, ResNetV2-50, MobileNetV1, and MobileNetV2 with transfer learning. To implement each pre-trained CNN architecture, we deployed the corresponded feature vector available from the TensorFlowHub, integrating it with dropout and dense layers to form a complete CNN model. Our findings indicated that the EfficientNetV2B0-21k (0.72B Floating-Point Operations and 7.1 M parameters) outperformed other CNN models in the IDC grading task. Nevertheless, we discovered that practically all selected CNN models perform well in the IDC grading task, with an average balanced accuracy of 0.936 ± 0.0189 on the cross-validation set and 0.9308 ± 0.0211on the test set.