A Novel Deep Learning Approach for Accurate Cancer Type and Subtype Identification

IEEE Access, Год журнала: 2024, Номер 12, С. 94116 - 94134

Опубликована: Янв. 1, 2024

Cancer is a disease where abnormal cells grow uncontrollably and spread to other body parts. It can originate anywhere in the human body, which consists of trillions cells. These continually divide, replenishing body's needs. As age or sustain damage, they naturally undergo apoptosis, allowing new take their place. Our research uses secondary dataset from Kaggle, comprising over 130,000 images representing various cancer types. We have developed novel Deep-learning model capable detecting classifying at early stages with remarkable accuracy. The classifies eight primary types 26 subtypes, each represented by 5,000 images. approach combines computational tools, including pre-trained Convolutional Neural Networks, Machine learning, Deep learning classifiers such as KNN SVM, innovative multimodal architectures merged CNN-LSTM hybrids. applied two distinct classification strategies. In our first approach, main class subclass are classified together. second predicts classes then subclasses concerning classification, achieved higher accuracy for Lymphoma than CNNs. Finally, X-OR gate-based fusion technique after prediction significantly reduces misclassifications enhances certainty findings reveal great levels 99.25% classifications 97.80% classifications. introduction models, Vception (VGG + Inception) Vmobilnet MobileNet), integrated LSTM, further advances diagnostic capabilities. Again, By utilizing an gate post-prediction Vmobilenet we 99.95% 99.13%, boosting confidence. Moreover, individually, 97.14% using PCA. This study not only sets benchmark detection but also promises improve patient care treatment outcomes significantly.

Язык: Английский

---

MediGPT: Exploring Potentials of Conventional and Large Language Models on Medical Data

IEEE Access, Год журнала: 2024, Номер 12, С. 103473 - 103487

Опубликована: Янв. 1, 2024

Medical text classification organizes medical documents into categories to streamline information retrieval and support clinical decision-making. Traditional machine learning techniques, including pre-trained language models, are effective but require extensive domain-specific training data, often underperform across languages, costly complex deploy on a large scale. In this study, we employed four datasets: Clinical trials cancer, encompassing 6 million statements from interventional cancer trial protocols; the Illness-dataset, consisting of 22,660 categorized tweets 2018 2019; Multi-View active for short in user-generated an extended version Illness-dataset same period; Symptom2Disease dataset, containing 1,200 data points used predict diseases based symptom descriptions. This study uses ChatGPT, particularly its ChatGPT-3.5 ChatGPT-4 versions, as viable alternative classifying texts. We investigate essential aspects, construction prompts, parsing responses, various strategic use GPT models optimize outcomes. Through comparative analysis with established methods like model fine-tuning prompt-tuning, our findings indicate that ChatGPT addresses these challenges efficiently matches performance traditional methods. Furthermore, enhanced capabilities proposed MediGPT (Medical Generative Pre-Trained Transformers) have led improvements 14.3%, 22.3%, 13.6%, 13.7% datasets, highlighting adaptability robustness diverse scenarios without need specialized domain adjustments. research underscores capability facilitate versatile AI framework processing, which could revolutionize informatics practices.

Язык: Английский

---

A Deep Dive into Vision Models: Comparing CNN and Transformer-Based Approaches for Pneumonia Detection

Опубликована: Фев. 18, 2025

Язык: Английский

---

Rapid diagnosis of IgA nephropathy using dual branch network combined with FTIR based on cross attention mechanism

Microchemical Journal, Год журнала: 2025, Номер unknown, С. 113762 - 113762

Опубликована: Апрель 1, 2025

Язык: Английский

---

GastroSmart: Precision GI Health Monitoring with Non-Invasive GMR

International Research Journal of Multidisciplinary Technovation, Год журнала: 2024, Номер unknown, С. 17 - 37

Опубликована: Окт. 30, 2024

Pathological conditions affecting the gastroenterological tract such as GERD, gastroparesis, gastric cancer, type 2 diabetes, and obesity among others present alarming levels of health risks. Conventional imaging methods ultrasonic have a very high cost do not provide real-time monitoring. To overcome these challenges, we new system based on GMR sensor capable non-invasively measuring volume over prolonged periods time. This uses Rational Dilation Wavelet Transformation in order to enhance accuracy evaluated dynamics. With help polynomial regression, changes can be predicted accurately by our model, which makes it possible prevent exacerbation gastrointestinal diseases early stages. The continuous evaluation condition patients their physical activity performed this non-invasive method will allow individualized treatment each patient best way improve healing without sacrificing safety. investigation is response for implementing low-cost effective solutions constant monitoring with distresses direction preventive nursing clinical care patients.

Язык: Английский

---

A two-phase transfer learning framework for gastrointestinal diseases classification

PeerJ Computer Science, Год журнала: 2024, Номер 10, С. e2587 - e2587

Опубликована: Дек. 19, 2024

Gastrointestinal (GI) disorders are common and often debilitating health issues that affect a significant portion of the population. Recent advancements in artificial intelligence, particularly computer vision algorithms, have shown great potential detecting classifying medical images. These algorithms utilize deep convolutional neural network architectures to learn complex spatial features images make predictions for similar unseen The proposed study aims assist gastroenterologists making more efficient accurate diagnoses GI patients by utilizing its two-phase transfer learning framework identify diseases from endoscopic Three pre-trained image classification models, namely Xception, InceptionResNetV2, VGG16, fine-tuned on publicly available datasets annotated tract. Additionally, two custom networks constructed fully trained comparative analysis their performance. Four different tasks examined based categories. architecture employing InceptionResNetV2 achieves most consistent generalized performance across tasks, yielding accuracy scores 85.7% general tract (eight-category classification), 97.6% three-diseases classification, 99.5% polyp identification (binary 74.2% binary esophagitis severity results indicate effectiveness clinical use enhance diseases, aiding early diagnosis treatment.

Язык: Английский

---

EMG-TransNN-MHA: A Transformer-Based Model for Enhanced Motor Intent Recognition in Assistive Robotics

2021 IEEE International Conference on Big Data (Big Data), Год журнала: 2024, Номер unknown, С. 8691 - 8693

Опубликована: Дек. 15, 2024

Язык: Английский

---