Developing a classification system for brain tumors using the ResNet152V2 CNN model architecture

Journal of Soft Computing Exploration, Journal Year: 2024, Volume and Issue: 5(2), P. 173 - 182

Published: June 21, 2024

According to The American Cancer Society, in 2021 there were 24,530 cases of brain and nervous system tumors. National Institute reports that are approximately 4.4 new tumors per 100,000 men women year. Brain can be detected using magnetic resonance imaging (MRI), a scanning tool uses field computer record images is able provide clear visualization differences soft tissue such as white matter gray matter. However, this cannot done optimally because it still relies on manual analysis, so classify tumor types larger datasets with the potential for error low level accuracy. To accurately determine type tumor, better classification method needed. aim study accuracy calcification deep learning model. In study, was carried out ResNet152V2 convolutional neural network (CNN) model which has depth 152 layers. dataset used 7,023 MRI consisting 1,645 meningiomas, 1,621 gliomas, 1,757 pituitary 2,000 normal. Research results show an value 94.44%, concluded performs well classifying medium physicians more diagnose patients accurately.

Language: Английский

---

Identification of Anomalies in Lung and Colon Cancer Using Computer Vision-Based Swin Transformer with Ensemble Model on Histopathological Images

Bioengineering, Journal Year: 2024, Volume and Issue: 11(10), P. 978 - 978

Published: Sept. 28, 2024

Lung and colon cancer (LCC) is a dominant life-threatening disease that needs timely attention precise diagnosis for efficient treatment. The conventional diagnostic techniques LCC regularly encounter constraints in terms of efficiency accuracy, thus causing challenges primary recognition Early the can immensely reduce probability death. In medical practice, histopathological study tissue samples generally uses classical model. Still, automated devices exploit artificial intelligence (AI) produce results diagnosis. histopathology, both machine learning (ML) deep (DL) approaches be deployed owing to their latent ability analyzing predicting physically accurate molecular phenotypes microsatellite uncertainty. this background, presents novel technique called Colon Cancer using Swin Transformer with an Ensemble Model on Histopathological Images (LCCST-EMHI). proposed LCCST-EMHI method focuses designing DL model classification images (HI). order achieve this, utilizes bilateral filtering (BF) get rid noise. Further, (ST) also employed purpose feature extraction. For detection process, ensemble classifier used three techniques: bidirectional long short-term memory multi-head (BiLSTM-MHA), Double Deep Q-Network (DDQN), sparse stacked autoencoder (SSAE). Eventually, hyperparameter selection models implemented utilizing walrus optimization algorithm (WaOA) method. illustrate promising performance approach, extensive range simulation analyses was conducted benchmark dataset. experimentation demonstrated approach over other recent methods.

Language: Английский

---

Exploring vision transformers and XGBoost as deep learning ensembles for transforming carcinoma recognition

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Dec. 3, 2024

Early detection of colorectal carcinoma (CRC), one the most prevalent forms cancer worldwide, significantly enhances prognosis patients. This research presents a new method for improving CRC using deep learning ensemble with Computer Aided Diagnosis (CADx). The involves combining pre-trained convolutional neural network (CNN) models, such as ADaRDEV2I-22, DaRD-22, and ADaDR-22, Vision Transformers (ViT) XGBoost. study addresses challenges associated imbalanced datasets necessity sophisticated feature extraction in medical image analysis. Initially, CKHK-22 dataset comprised 24 classes. However, we refined it to 14 classes, which led an improvement data balance quality. enabled more precise improved classification results. We created two models: first model used capture long-range spatial relationships images, while second combined CNNs XGBoost facilitate structured classification. implemented DCGAN-based augmentation enhance dataset's diversity. tests showed big improvements performance, ADaDR-22 + Transformer group getting best results, testing accuracy 93.4% AUC 98.8%. In contrast, had 97.8% 92.2%. These findings highlight efficacy proposed models detecting importance well-balanced, high-quality datasets. clinical diagnostic capabilities analysis or early detection.

Language: Английский

---

A Review on Lung and Colon Combine Cancer Detection using ML and DL Techniques

International Journal of Scientific Research in Computer Science Engineering and Information Technology, Journal Year: 2024, Volume and Issue: 10(5), P. 24 - 35

Published: Sept. 5, 2024

The detection of lung and colon cancer is a critical challenge in medical diagnosis, machine learning (ML) deep (DL) techniques are increasingly being used to enhance accuracy efficiency. This review focuses on the integration ML DL methods for combined cancer, emphasizing their strengths, limitations, future potential. motivation behind this study address growing demand accurate early these cancers, which significantly impacts treatment outcomes. Current often struggle with feature complexity, image variability, computational intensity, limit real-world applicability. aim consolidate various that have been employed purpose, highlighting how hybrid models can improve rates. objective provide comprehensive analysis different methodologies, datasets, pre-processing techniques, extraction methods, evaluation parameters. also explores recent advancements, such as transfer fine-tuning further optimize performance detection. findings suggest while current show promise, improvements model generalization, interpretability, efficiency required overcome existing limitations expand clinical use.

Language: Английский

---

Transfer Learning Method Towards Lung and Colon Cancers Automated Analysis in Histopathological Images

Published: Dec. 3, 2024

Language: Английский

---