Wildfires pose a severe threat to ecological systems, human life, and infrastructure, making early detection critical for timely intervention. Traditional fire detection systems rely heavily on single-sensor approaches and are often hindered by environmental conditions such as smoke, fog, or nighttime scenarios. This paper proposes Adaptive Multi-Sensor Oriented Object Detection with Space–Frequency Selective Convolution (AMSO-SFS), a novel deep learning-based model optimized for drone-based wildfire and smoke detection. AMSO-SFS combines optical, infrared, and Synthetic Aperture Radar (SAR) data to detect fire and smoke under varied visibility conditions. The model introduces a Space–Frequency Selective Convolution (SFS-Conv) module to enhance the discriminative capacity of features in both spatial and frequency domains. Furthermore, AMSO-SFS utilizes weakly supervised learning and adaptive scale and angle detection to identify fire and smoke regions with minimal labeled data. Extensive experiments show that the proposed model outperforms current state-of-the-art (SoTA) models, achieving robust detection performance while maintaining computational efficiency, making it suitable for real-time drone deployment.

The present-day issues related to the cotton-growing industry, namely yield estimation, pest effect, and growth phase diagnostics, call for integrated, scalable monitoring solutions. This write-up reveals Cotton Multitask Learning (CMTL), a transformer-driven multitask framework that launches three major agronomic tasks from UAV pictures at one go: boll detection, pest damage segmentation, and phenological stage classification. CMTL does not change separate pipelines, but rather merges these goals using a Cross-Level Multi-Granular Encoder (CLMGE) and a Multitask Self-Distilled Attention Fusion (MSDAF) module that both allow mutual learning across tasks and still keep their specific features. The biologically guided Stage Consistency Loss is the part of the architecture of the network that enables the system to carry out growth stage transitions that occur in reality. We executed CMTL on a tri-source UAV dataset that fused over 2100 labeled images from public and private collections, representing a variety of crop stages and conditions. The model showed its virtues state-of-the-art baselines in all the tasks: setting 0.913 mAP for boll detection, 0.832 IoU for pest segmentation, and 0.936 accuracy for growth stage classification. Additionally, it runs at the fastest speed of performance on edge devices such as NVIDIA Jetson Xavier NX (Manufactured in Shanghai, China), which makes it ideal for deployment. These outcomes evoke CMTL’s promise as a single and productive instrument of aerial crop intelligence in precision cotton agriculture.

The timely and accurate detection of brain tumors is crucial for effective medical intervention, especially in resource-constrained settings. This study proposes a lightweight and efficient RetinaNet variant tailored for medical edge device deployment. The model reduces computational overhead while maintaining high detection accuracy by replacing the computationally intensive ResNet backbone with MobileNet and leveraging depthwise separable convolutions. The modified RetinaNet achieves an average precision (AP) of 32.1, surpassing state-of-the-art models in small tumor detection (AP_S: 14.3) and large tumor localization (AP_L: 49.7). Furthermore, the model significantly reduces computational costs, making real-time analysis feasible on low-power hardware. Clinical relevance is a key focus of this work. The proposed model addresses the diagnostic challenges of small, variable-sized tumors often overlooked by existing methods. Its lightweight architecture enables accurate and timely tumor localization on portable devices, bridging the gap in diagnostic accessibility for underserved regions. Extensive experiments on the BRATS dataset demonstrate the model robustness across tumor sizes and configurations, with confidence scores consistently exceeding 81%. This advancement holds the potential for improving early tumor detection, particularly in remote areas lacking advanced medical infrastructure, thereby contributing to better patient outcomes and broader accessibility to AI-driven diagnostic tools.

This paper presents an enhanced approach to kidney segmentation using a modified CLAHE preprocessing method, aimed at improving image clarity and CNN performance on the KiTS19 dataset. To assess the impact of the modified CLAHE method, we conducted quality evaluations using the BRISQUE metric, comparing the original, standard CLAHE and modified CLAHE versions of the dataset. The BRISQUE score decreased from 28.8 in the original dataset to 21.1 with the modified CLAHE method, indicating a significant improvement in image quality. Furthermore, CNN segmentation accuracy rose from 0.951 with the original dataset to 0.996 with the modified CLAHE method, outperforming the accuracy achieved with standard CLAHE preprocessing (0.969). These results highlight the benefits of the modified CLAHE method in refining image quality and enhancing segmentation performance. This study highlights the value of adaptive preprocessing in medical imaging workflows and shows that CNN-based kidney segmentation accuracy may be greatly increased by altering conventional CLAHE. Our method provides insightful information on optimizing preprocessing for medical imaging applications, leading to more accurate and dependable segmentation results for better clinical diagnosis.

Pneumonia is a form of acute respiratory infection affecting the lungs. Symptoms of viral and bacterial pneumonia are similar. Rapid diagnosis of the disease is difficult, since polymerase chain reaction-based methods, which have the greatest reliability, provide results in a few hours, while ensuring high requirements for compliance with the analysis technology and professionalism of the personnel. This study proposed a Concatenated CNN model for pneumonia detection combined with a fuzzy logic-based image improvement method. The fuzzy logic-based image enhancement process is based on a new fuzzification refinement algorithm, with significantly improved image quality and feature extraction for the CCNN model. Four datasets, original and upgraded images utilizing fuzzy entropy, standard deviation, and histogram equalization, were utilized to train the algorithm. The CCNN’s performance was demonstrated to be significantly improved by the upgraded datasets, with the fuzzy entropy-added dataset producing the best results. The suggested CCNN attained remarkable classification metrics, including 98.9% accuracy, 99.3% precision, 99.8% F1-score, and 99.6% recall. Experimental comparisons showed that the fuzzy logic-based enhancement worked significantly better than traditional image enhancement methods, resulting in higher diagnostic precision. This study demonstrates how well deep learning models and sophisticated image enhancement techniques work together to analyze medical images.

In this study, a novel method combining contour analysis with deep CNN is applied for fire detection. The method was made for fire detection using two main algorithms: one which detects the color properties of the fires, and another which analyzes the shape through contour detection. To overcome the disadvantages of previous methods, we generate a new labeled dataset, which consists of small fire instances and complex scenarios. We elaborated the dataset by selecting regions of interest (ROI) for enhanced fictional small fires and complex environment traits extracted through color characteristics and contour analysis, to better train our model regarding those more intricate features. Results of the experiment showed that our improved CNN model outperformed other networks. The accuracy, precision, recall and F1 score were 99.4%, 99.3%, 99.4% and 99.5%, respectively. The performance of our new approach is enhanced in all metrics compared to the previous CNN model with an accuracy of 99.4%. In addition, our approach beats many other state-of-the-art methods as well: Dilated CNNs (98.1% accuracy), Faster R-CNN (97.8% accuracy) and ResNet (94.3%). This result suggests that the approach can be beneficial for a variety of safety and security applications ranging from home, business to industrial and outdoor settings.