Self-attention-based generative adversarial network optimized with color harmony algorithm for brain tumor classification

Electromagnetic Biology and Medicine, Journal Year: 2024, Volume and Issue: 43(1-2), P. 31 - 45

Published: Feb. 18, 2024

This paper proposes a novel approach, BTC-SAGAN-CHA-MRI, for the classification of brain tumors using SAGAN optimized with Color Harmony Algorithm. Brain cancer, its high fatality rate worldwide, especially in case tumors, necessitates more accurate and efficient methods. While existing deep learning approaches tumor have been suggested, they often lack precision require substantial computational time.The proposed method begins by gathering input MR images from BRATS dataset, followed pre-processing step Mean Curvature Flow-based approach to eliminate noise. The pre-processed then undergo Improved Non-Sub sampled Shearlet Transform (INSST) extracting radiomic features. These features are fed into SAGAN, which is Algorithm categorize different types, including Gliomas, Meningioma, Pituitary tumors. innovative shows promise enhancing efficiency classification, holding potential improved diagnostic outcomes field medical imaging. accuracy acquired identification 99.29%. BTC-SAGAN-CHA-MRI technique achieves 18.29%, 14.09% 7.34% higher 67.92%,54.04%, 59.08% less Computation Time when analyzed models, like diagnosis utilizing convolutional neural network transfer (BTC-KNN-SVM-MRI); M3BTCNet: multi model categorization under metaheuristic optimization (BTC-CNN-DEMFOA-MRI), depending upon hierarchical classifier tumour (BTC-Hie DNN-MRI) respectively.

Language: Английский

---

PRIS: Practical robust invertible network for image steganography

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 133, P. 108419 - 108419

Published: April 11, 2024

Language: Английский

---

Mobile robot path planning based on bi-population particle swarm optimization with random perturbation strategy

Journal of King Saud University - Computer and Information Sciences, Journal Year: 2024, Volume and Issue: 36(2), P. 101974 - 101974

Published: Feb. 1, 2024

Path planning for mobile robots poses a challenging optimization problem, requiring the discovery of near-optimal path within diverse constraints. Conventional particle swarm (PSO) algorithms encounter limitations in solving constrained problems, vulnerability to local optima, and premature convergence. To address these challenges, this paper proposes bi-population PSO algorithm with random perturbation strategy (BPPSO), which divides particles into two subpopulations. The first subpopulation enhances global search capabilities by considering quality optimal solution randomly selected when updating velocities. second strengthens using linear cognitive coefficient adjustment strategy. Moreover, counter tracks iteration without improvement best position. Upon reaching predefined threshold, is added positions all both subpopulations, increasing diversity enhancing ability escape optima. performance BPPSO was experimentally validated across three benchmark functions four environment models. results have demonstrated that proposed outperforms existing other established terms running time, highlighting feasibility resolving challenge robot planning.

Language: Английский

---

Performance analysis of deep learning-based object detection algorithms on COCO benchmark: a comparative study

Journal of Engineering and Applied Science, Journal Year: 2024, Volume and Issue: 71(1)

Published: March 21, 2024

Abstract This paper thoroughly explores the role of object detection in smart cities, specifically focusing on advancements deep learning-based methods. Deep learning models gain popularity for their autonomous feature learning, surpassing traditional approaches. Despite progress, challenges remain, such as achieving high accuracy urban scenes and meeting real-time requirements. The study aims to contribute by analyzing state-of-the-art algorithms, identifying accurate evaluating performance using Average Precision at Medium Intersection over Union (IoU) metric. reported results showcase various algorithms’ performance, with Dynamic Head (DyHead) emerging top scorer, excelling accurately localizing classifying objects. Its precision recall medium IoU thresholds signify robustness. suggests considering mean (mAP) metric a comprehensive evaluation across thresholds, if available. this, DyHead stands out superior algorithm, particularly making it suitable precise city applications. analysis is reinforced Low (APL), consistently depicting DyHead’s superiority. These findings provide valuable insights researchers practitioners, guiding them toward employing tasks prioritizing localization classification cities. Overall, navigates through complexities environments, presenting leading solution robust metrics.

Language: Английский

---

Advancing cancer diagnosis and treatment: integrating image analysis and AI algorithms for enhanced clinical practice

Artificial Intelligence Review, Journal Year: 2025, Volume and Issue: 58(4)

Published: Jan. 25, 2025

Language: Английский

---

A Yolo-based Approach for Fire and Smoke Detection in IoT Surveillance Systems

International Journal of Advanced Computer Science and Applications, Journal Year: 2024, Volume and Issue: 15(1)

Published: Jan. 1, 2024

Fire and smoke detection in IoT surveillance systems is of utmost importance for ensuring public safety preventing property damage. While traditional methods have been used fire detection, deep learning-based approaches gained significant attention due to their ability learn complex patterns achieve high accuracy. This paper addresses the current research challenge achieving accuracy rates with while keeping computation costs low. proposes a method based on Yolov8 algorithm that effectively tackles this through model generation using custom dataset model's training, validation, testing. The efficacy succinctly assessed by precision, recall F1-curve metrics, notable proficiency crucial early warnings prevention. Experimental results performance evaluations show our proposed outperforms other state-of-the-art methods. makes it promising approach systems.

Language: Английский

---

EfficientUNetViT: Efficient Breast Tumor Segmentation utilizing U-Net Architecture and Pretrained Vision Transformer

Published: Aug. 14, 2024

This study introduces a sophisticated neural network structure for segmenting breast tumors. It achieves this by combining pretrained Vision Transformer (ViT) model with U-Net framework. The architecture, commonly employed biomedical image segmentation, is further enhanced Depthwise Separable Convolutional Blocks to decrease computational complexity and parameter count, resulting in better efficiency less overfitting. Transformer, renowned its robust feature extraction capabilities utilizing self-attention processes, efficiently captures the overall context within images, surpassing performance of conventional convolutional networks. By using ViT as encoder our model, we take advantage extensive representations acquired from datasets, major enhancement model’s ability generalize train efficiently. suggested has exceptional cancers medical highlighting advantages integrating transformer-based encoders efficient topologies. hybrid methodology emphasizes transformers field processing establishes new standard accuracy activities related tumor segmentation.

Language: Английский

---

EfficientUNetViT: Efficient Breast Tumor Segmentation Utilizing UNet Architecture and Pretrained Vision Transformer

Bioengineering, Journal Year: 2024, Volume and Issue: 11(9), P. 945 - 945

Published: Sept. 21, 2024

This study introduces a sophisticated neural network structure for segmenting breast tumors. It achieves this by combining pretrained Vision Transformer (ViT) model with UNet framework. The architecture, commonly employed biomedical image segmentation, is further enhanced depthwise separable convolutional blocks to decrease computational complexity and parameter count, resulting in better efficiency less overfitting. ViT, renowned its robust feature extraction capabilities utilizing self-attention processes, efficiently captures the overall context within images, surpassing performance of conventional networks. By using ViT as encoder our model, we take advantage extensive representations acquired from datasets, major enhancement model’s ability generalize train efficiently. suggested has exceptional cancers medical highlighting advantages integrating transformer-based encoders efficient topologies. hybrid methodology emphasizes transformers field processing establishes new standard accuracy activities related tumor segmentation.

Language: Английский

---

The segmentation of nanoparticles with a novel approach of HRU2-Net†

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Jan. 16, 2025

Nanoparticles have great potential for the application in new energy and aerospace fields. The distribution of nanoparticle sizes is a critical determinant material properties serves as significant parameter defining characteristics zero-dimensional nanomaterials. In this study, we proposed HRU2-Net†, an enhancement U2-Net† model, featuring multi-level semantic information fusion. This approach exhibits strong competitiveness refined segmentation capabilities segmentation. It achieves Mean intersection over union (MIoU) 87.31%, with accuracy rate exceeding 97.31%, leading to improvement effectiveness precision. results show that deep learning-based method significantly enhances efficacy nanomaterial research, which holds substantial significance advancement science.

Language: Английский

---

SDC-HSDD-NDSA: Structure Detecting Cluster by Hierarchical Secondary Directed Differential with Normalized Density and Self-Adaption

Information Sciences, Journal Year: 2025, Volume and Issue: unknown, P. 121916 - 121916

Published: Jan. 1, 2025

Language: Английский

---