CG‐Net: A novel CNN framework for gastrointestinal tract diseases classification

International Journal of Imaging Systems and Technology, Journal Year: 2024, Volume and Issue: 34(3)

Published: April 17, 2024

Abstract The classification of medical images has had a significant influence on the diagnostic techniques and therapeutic interventions. Conventional disease diagnosis procedures require substantial amount time effort to accurately diagnose. Based global statistics, gastrointestinal cancer been recognized as major contributor cancer‐related deaths. complexities involved in resolving tract (GIT) ailments arise from need for elaborate methods precisely identify exact location problem. Therefore, doctors frequently use wireless capsule endoscopy diagnose treat GIT problems. This research aims develop robust framework using deep learning effectively classify diseases purposes. A CNN based framework, conjunction with feature selection method, proposed improve rate. evaluated various performance measures, including accuracy, recall, precision, F1 measure, mean absolute error, squared error.

Language: Английский

---

Uncovering the Potential of Indoor Localization: Role of Deep and Transfer Learning

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 73980 - 74010

Published: Jan. 1, 2024

Indoor localization (IL) is a significant topic of study with several practical applications, particularly in the context Internet Things (IoT) and smart cities. The area IL has evolved greatly recent years due to introduction numerous technologies such as WiFi, Bluetooth, cameras, other sensors. Despite growing interest this field, there are challenges drawbacks that must be addressed develop more accurate sustainable systems for IL. This review gives an in-depth look into IL, covering most promising artificial intelligence-based hybrid strategies have shown excellent potential overcoming some limitations classic methods within IoT environments. In addition, paper investigates significance high-quality datasets evaluation metrics design assessment algorithms. Furthermore, overview emphasizes crucial role machine learning techniques, deep transfer learning, play advancement A focus on importance various technologies, methods, techniques being used improve it. Finally, survey highlights need continued research development create scalable can applied across range IoT-related industries, evacuation-egress routes, hazard-crime detection, occupancy-driven energy reduction asset tracking management.

Language: Английский

---

Federated and transfer learning for cancer detection based on image analysis

Neural Computing and Applications, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 10, 2025

Language: Английский

---

ViTs as backbones: Leveraging vision transformers for feature extraction

Information Fusion, Journal Year: 2025, Volume and Issue: unknown, P. 102951 - 102951

Published: Jan. 1, 2025

Language: Английский

---

Exploring the application of knowledge transfer to sports video data

Frontiers in Sports and Active Living, Journal Year: 2025, Volume and Issue: 6

Published: Feb. 7, 2025

The application of Artificial Intelligence (AI) and Computer Vision (CV) in sports has generated significant interest enhancing viewer experience through graphical overlays predictive analytics, as well providing valuable insights to coaches. However, more efficient methods are needed that can be applied across different without incurring high data annotation or model training costs. A major limitation deep learning models on large datasets is the resource requirement for reproducing results. Transfer Learning Zero-Shot (ZSL) offer promising alternatives this approach. For example, ZSL player re-identification (a crucial step complex behavioral analysis) involves re-identifying players videos having seen examples those during phase. This study investigates performance various techniques context Rugby League Netball. We focus use feature embeddings measure similarity between players. To support our experiments, we created two comprehensive broadcast video clips: one with nearly 35,000 frames another close 14,000 Netball, each annotated IDs actions. Our approach leverages pre-trained extract evaluation under a challenging testing environmnet. Results demonstrate outperformed general person datasets. Part-based showed particular promise handling challenges dynamic environments, while non-part-based struggled due background interference.

Language: Английский

---

A transfer learning-based YOLO network for sewer defect detection in comparison to classic object detection methods

Developments in the Built Environment, Journal Year: 2023, Volume and Issue: 15, P. 100191 - 100191

Published: July 1, 2023

Deep learning has shown promising performance in automated sewer defect detection, however, is generally data-driven and computationally intensive. Transfer (TL) solves the problem of data limitations avoids need to build models from scratch. This study compared a TL-based YOLO network (with 11 pretrained backbone CNNs) with four mainstream object detection methods (ODMs) for detecting five types defects. Results showed that transferred outperformed other ODMs, improved precision, computation speed intersection over union (IoU). Among CNNs, Resnet18 achieved best performance, while Inceptionresnetv2 was least effective. The ODMs worked disjoint, whereas tree root crack were most challenging predict. work not only illustrated benefits TL, but also provided technical guidance practitioners who lack expertise rely on TL better detection.

Language: Английский

---

Artificial Intelligence for Cochlear Implants: Review of Strategies, Challenges, and Perspectives

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 119015 - 119038

Published: Jan. 1, 2024

Automatic speech recognition (ASR) plays a pivotal role in our daily lives, offering utility not only for interacting with machines but also facilitating communication individuals partial or profound hearing impairments. The process involves receiving the signal analog form, followed by various processing algorithms to make it compatible devices of limited capacities, such as cochlear implants (CIs). Unfortunately, these implants, equipped finite number electrodes, often result distortion during synthesis. Despite efforts researchers enhance received quality using state-of-the-art techniques, challenges persist, especially scenarios involving multiple sources speech, environmental noise, and other adverse conditions. advent new artificial intelligence (AI) methods has ushered cutting-edge strategies address limitations difficulties associated traditional techniques dedicated CIs. This review aims comprehensively cover advancements CI-based ASR enhancement, among related aspects. primary objective is provide thorough overview metrics datasets, exploring capabilities AI this biomedical field, summarizing commenting on best results obtained. Additionally, will delve into potential applications suggest future directions bridge existing research gaps domain.

Language: Английский

---

Artificial intelligence-based video monitoring of movement disorders in the elderly: a review on current and future landscapes

Singapore Medical Journal, Journal Year: 2024, Volume and Issue: 65(3), P. 141 - 149

Published: March 1, 2024

Abstract Due to global ageing, the burden of chronic movement and neurological disorders (Parkinson’s disease essential tremor) is rapidly increasing. Current diagnosis monitoring these rely largely on face-to-face assessments utilising clinical rating scales, which are semi-subjective time-consuming. To address challenges, utilisation artificial intelligence (AI) has emerged. This review explores advantages challenges associated with using AI-driven video care for elderly patients disorders. The AI-based systems offer improved efficiency objectivity in remote patient monitoring, enabling real-time analysis data, more uniform outcomes augmented support trials. However, such as quality, privacy compliance noisy training labels, during development need be addressed. Ultimately, advancement expected evolve towards discreet, home-based evaluations routine daily activities. progression must incorporate data security, ethical considerations adherence regulatory standards.

Language: Английский

---

Robustness in deep learning models for medical diagnostics: security and adversarial challenges towards robust AI applications

Artificial Intelligence Review, Journal Year: 2024, Volume and Issue: 58(1)

Published: Nov. 8, 2024

The current study investigates the robustness of deep learning models for accurate medical diagnosis systems with a specific focus on their ability to maintain performance in presence adversarial or noisy inputs. We examine factors that may influence model reliability, including complexity, training data quality, and hyperparameters; we also security concerns related attacks aim deceive along privacy seek extract sensitive information. Researchers have discussed various defenses these enhance robustness, such as input preprocessing, mechanisms like augmentation uncertainty estimation. Tools packages extend reliability features frameworks TensorFlow PyTorch are being explored evaluated. Existing evaluation metrics additionally This paper concludes by discussing limitations existing literature possible future research directions continue enhancing status this topic, particularly domain, ensuring AI trustworthy, reliable, stable.

Language: Английский

---

Fake news detection in Dravidian languages using transfer learning with adaptive finetuning

Engineering Applications of Artificial Intelligence, Journal Year: 2023, Volume and Issue: 126, P. 106877 - 106877

Published: Aug. 9, 2023

Language: Английский

---