Breast Cancer Detection on Histopathological Images Using a Composite Dilated Backbone Network

Computational Intelligence and Neuroscience, Journal Year: 2022, Volume and Issue: 2022, P. 1 - 10

Published: July 6, 2022

Breast cancer is a lethal illness that has high mortality rate. In treatment, the accuracy of diagnosis crucial. Machine learning and deep may be beneficial to doctors. The proposed backbone network critical for present performance CNN-based detectors. Integrating dilated convolution, ResNet, Alexnet increases detection performance. composite (CDBN) an innovative method integrating many identical backbones into single robust backbone. Hence, CDBN uses lead feature maps identify objects. It feeds high-level output features from previous next in stepwise way. We show most contemporary detectors can easily include improve achieved mAP improvements ranging 1.5 3.0 percent on breast histopathological image classification (BreakHis) dataset. Experiments have also shown instance segmentation improved. BreakHis dataset, enhances baseline detector cascade mask R-CNN (mAP = 53.3). does not need pretraining. creates traits by combining low-level elements. This made up several are linked together. considers CDBN.

Language: Английский

---

On the Analyses of Medical Images Using Traditional Machine Learning Techniques and Convolutional Neural Networks

Archives of Computational Methods in Engineering, Journal Year: 2023, Volume and Issue: 30(5), P. 3173 - 3233

Published: April 4, 2023

Convolutional neural network (CNN) has shown dissuasive accomplishment on different areas especially Object Detection, Segmentation, Reconstruction (2D and 3D), Information Retrieval, Medical Image Registration, Multi-lingual translation, Local language Processing, Anomaly Detection video Speech Recognition. CNN is a special type of Neural Network, which compelling effective learning ability to learn features at several steps during augmentation the data. Recently, interesting inspiring ideas Deep Learning (DL) such as activation functions, hyperparameter optimization, regularization, momentum loss functions improved performance, operation execution Different internal architecture innovation representational style significantly performance. This survey focuses taxonomy deep learning, models vonvolutional network, depth width in addition components, applications current challenges learning.

Language: Английский

---

The Role of Deep Learning in Advancing Breast Cancer Detection Using Different Imaging Modalities: A Systematic Review

Cancers, Journal Year: 2022, Volume and Issue: 14(21), P. 5334 - 5334

Published: Oct. 29, 2022

Breast cancer is among the most common and fatal diseases for women, no permanent treatment has been discovered. Thus, early detection a crucial step to control cure breast that can save lives of millions women. For example, in 2020, more than 65% patients were diagnosed an stage cancer, from which all survived. Although effective approach treatment, screening conducted by radiologists very expensive time-consuming. More importantly, conventional methods analyzing images suffer high false-detection rates. Different imaging modalities are used extract analyze key features affecting diagnosis cancer. These be divided into subgroups such as mammograms, ultrasound, magnetic resonance imaging, histopathological images, or any combination them. Radiologists pathologists produced these manually, leads increase risk wrong decisions detection. utilization new automatic kinds assist interpret required. Recently, artificial intelligence (AI) widely utilized automatically improve different types specifically thereby enhancing survival chance patients. Advances AI algorithms, deep learning, availability datasets obtained various have opened opportunity surpass limitations current analysis methods. In this article, we first review modalities, their strengths limitations. Then, explore summarize recent studies employed using modalities. addition, report available on breast-cancer important developing AI-based algorithms training learning models. conclusion, paper tries provide comprehensive resource help researchers working analysis.

Language: Английский

---

Classification of Breast Cancer Histopathological Images Using DenseNet and Transfer Learning

Computational Intelligence and Neuroscience, Journal Year: 2022, Volume and Issue: 2022, P. 1 - 31

Published: Oct. 10, 2022

Breast cancer is one of the most common invading cancers in women. Analyzing breast nontrivial and may lead to disagreements among experts. Although deep learning methods achieved an excellent performance classification tasks including histopathological images, existing state-of-the-art are computationally expensive overfit due extracting features from in-distribution images. In this paper, our contribution mainly twofold. First, we perform a short survey on deep-learning-based models for classifying images investigate popular optimized training-testing ratios. Our findings reveal that ratio image 70%: 30%, whereas best (e.g., accuracy) by using 80%: 20% identical dataset. Second, propose method named DenTnet classify chiefly. utilizes principle transfer solve problem same distribution DenseNet as backbone model. The proposed shown be superior comparison number leading terms detection accuracy (up 99.28% BreaKHis dataset deeming 20%) with good generalization ability computational speed. limitation requirement high computation utilization feature mitigated dint DenTnet.

Language: Английский

---

Multiple Types of Cancer Classification Using CT/MRI Images Based on Learning Without Forgetting Powered Deep Learning Models

IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 10336 - 10354

Published: Jan. 1, 2023

Cancer is the second biggest cause of death worldwide, accounting for one every six deaths. On other hand, early detection disease significantly improves chances survival. The use Artificial Intelligence (AI) to automate cancer might allow us evaluate more cases in less time. In this research, AI-based deep learning models are proposed classify images eight kinds cancer, such as lung, brain, breast, and cervical cancer. This work evaluates models, namely Convolutional Neural Networks (CNN), against classifying with traits. Pre-trained CNN variants MobileNet, VGGNet, DenseNet employed transfer knowledge they learned ImageNet dataset detect different cells. We Bayesian Optimization find suitable values hyperparameters. However, could make it so that can no longer datasets were initially trained. So, we Learning without Forgetting (LwF), which trains network using only new task data while keeping network’s original abilities. results experiments show based on accurate than current state-of-the-art techniques. also LwF better both have been trained before.

Language: Английский

---

Advancing cancer diagnosis and prognostication through deep learning mastery in breast, colon, and lung histopathology with ResoMergeNet

Computers in Biology and Medicine, Journal Year: 2024, Volume and Issue: 185, P. 109494 - 109494

Published: Dec. 4, 2024

Language: Английский

---

Chaotic Sparrow Search Algorithm with Deep Transfer Learning Enabled Breast Cancer Classification on Histopathological Images

Cancers, Journal Year: 2022, Volume and Issue: 14(11), P. 2770 - 2770

Published: June 2, 2022

Breast cancer is the major cause behind death of women worldwide and responsible for several deaths each year. Even though there are means to identify breast cancer, histopathological diagnosis now considered gold standard in cancer. However, difficulty image rapid rise workload render this process time-consuming, outcomes might be subjected pathologists' subjectivity. Hence, development a precise automatic analysis method essential field. Recently, deep learning pathological classification has made significant progress, which become mainstream This study introduces novel chaotic sparrow search algorithm with transfer learning-enabled (CSSADTL-BCC) model on images. The presented CSSADTL-BCC mainly focused recognition To accomplish this, primarily applies Gaussian filtering (GF) approach eradicate occurrence noise. In addition, MixNet-based feature extraction employed generate useful set vectors. Moreover, stacked gated recurrent unit (SGRU) exploited allot class labels. Furthermore, CSSA applied optimally modify hyperparameters involved SGRU model. None earlier works have utilized hyperparameter-tuned HIs. design optimal hyperparameter tuning demonstrates novelty work. performance validation tested by benchmark dataset, results reported superior execution over recent state-of-the-art approaches.

Language: Английский

---

Automatic Blob Detection Method for Cancerous Lesions in Unsupervised Breast Histology Images

Bioengineering, Journal Year: 2025, Volume and Issue: 12(4), P. 364 - 364

Published: March 31, 2025

The early detection of cancerous lesions is a challenging task given the cancer biology and variability in tissue characteristics, thus rendering medical image analysis tedious time-inefficient. In past, conventional computer-aided diagnosis (CAD) methods have heavily relied on visual inspection images, which ineffective, particularly for large visible such images. Additionally, face challenges analyzing objects images due to overlapping/intersecting inability resolve their boundaries/edges. Nevertheless, breast key determinant treatment. this study, we present deep learning-based technique lesion detection, namely blob automatically detects hidden inaccessible unsupervised human histology Initially, approach prepares pre-processes data through various augmentation increase dataset size. Secondly, stain normalization applied augmented separate nucleus features from structures. Thirdly, morphology operation techniques, erosion, dilation, opening, distance transform, are used enhance by highlighting foreground background pixels while removing overlapping regions highlighted image. Subsequently, segmentation handled via connected components method, groups pixel with similar intensity values assigns them relevant labeled (binary masks). These binary masks then active contours method further boundaries/edges ROIs. Finally, learning recurrent neural network (RNN) model extracts edges method. This proposed utilizes capabilities both limitations detection. evaluated 27,249 unsupervised, it shows significant evaluation result form 98.82% F1 accuracy score.

Language: Английский

---

An enhanced framework for detecting the histopathology cancer using machine learning technique

AIP conference proceedings, Journal Year: 2025, Volume and Issue: 3157, P. 080006 - 080006

Published: Jan. 1, 2025

Language: Английский

---

An Approach to Breast Cancer Detection with Histopathological Images Using Transfer Learning

Lecture notes in networks and systems, Journal Year: 2025, Volume and Issue: unknown, P. 322 - 331

Published: Jan. 1, 2025

Language: Английский

---