Classification of Breast Cancer using Histology images: Handcrafted and Pre-Trained Features Based Approach

Kundale, Jyoti; Dhage, Sudhir

doi:10.1088/1757-899x/1074/1/012008

Cited by 10 publications

(3 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The dataset is divided into training, validation, and testing sets to ensure proper evaluation of the models’ performance. It ensures that the same set of images is used for training, validation, and testing, with different DL models [ 26 , 27 , 28 ]. Six different DL models are employed for image classification, ranging from less complex to high complex DL models including MobileNet, XceptionNet, InceptionV3, ResNet50, VGG16, and VGG19, chosen based on their proven performance in image classification tasks and compatibility with the stained histopathological images [ 29 , 30 , 30 , 31 , 32 ].…”

Section: Methodsmentioning

confidence: 99%

Impact of H&E Stain Normalization on Deep Learning Models in Cancer Image Classification: Performance, Complexity, and Trade-Offs

Madusanka,

Jayalath,

Fernando

et al. 2023

Cancers

View full text Add to dashboard Cite

Accurate classification of cancer images plays a crucial role in diagnosis and treatment planning. Deep learning (DL) models have shown promise in achieving high accuracy, but their performance can be influenced by variations in Hematoxylin and Eosin (H&E) staining techniques. In this study, we investigate the impact of H&E stain normalization on the performance of DL models in cancer image classification. We evaluate the performance of VGG19, VGG16, ResNet50, MobileNet, Xception, and InceptionV3 on a dataset of H&E-stained cancer images. Our findings reveal that while VGG16 exhibits strong performance, VGG19 and ResNet50 demonstrate limitations in this context. Notably, stain normalization techniques significantly improve the performance of less complex models such as MobileNet and Xception. These models emerge as competitive alternatives with lower computational complexity and resource requirements and high computational efficiency. The results highlight the importance of optimizing less complex models through stain normalization to achieve accurate and reliable cancer image classification. This research holds tremendous potential for advancing the development of computationally efficient cancer classification systems, ultimately benefiting cancer diagnosis and treatment.

show abstract

Section: Methodsmentioning

confidence: 99%

Impact of H&E Stain Normalization on Deep Learning Models in Cancer Image Classification: Performance, Complexity, and Trade-Offs

Madusanka,

Jayalath,

Fernando

et al. 2023

Cancers

View full text Add to dashboard Cite

show abstract

“…This approach yielded accuracy improvements ranging from 17.14% to 25.30% compared to models trained from scratch. Similarly, Kundale and Dhage [29] employed pre-trained networks such as VGG, ResNet, and GoogLeNet for histopathology breast cancer classification, outperforming custom CNN models by margins of 19.48% to 29.05%. Celik et al [26] utilized pre-trained DenseNet161 and ResNet50 models for breast cancer classification, achieving balanced accuracy improvements of 1.96% to 6.73% compared to models in other state-of-the-art studies.…”

Section: Related Work a Conventional Transfer Learningmentioning

confidence: 99%

Histopathological Cancer Detection Using Intra-Domain Transfer Learning and Ensemble Learning

Yong,

Hum,

Lai

et al. 2024

IEEE Access

View full text Add to dashboard Cite

Cancer is a leading global cause of death. Histopathology image analysis is widely recognized as the gold standard for cancer diagnosis, playing a crucial role in early detection and reducing mortality rates. However, this diagnostic task is performed manually by pathologists, leading to high human errors and variabilities due to the huge number of images to screen and tissue complexities. With the emergence of deep learning, specifically the Convolutional Neural Network (CNN), in computer-aided diagnosis, there is a growing interest from the medical community to automate the labor-intensive manual image screening. Despite its promising performance, deep learning models still encounter challenges when it comes to extracting comprehensive histopathological features for optimal results. To tackle this issue, our study introduces deep learning models based on intra-domain transfer learning and ensemble learning. We evaluated these models on public histopathology datasets, including Gastric Histopathology Sub-size Image Database (GasHisSDB), Chaoyang colorectal, and Clinical Proteomic Tumor Analysis Consortium Clear Cell Renal Cell Carcinoma (CPTAC-CCRCC). Our models achieved state-of-the-art accuracy: 99.78% on GasHisSDB, 85.69% on Chaoyang, and 99.17% on CPTAC-CCRCC. These results highlight our models' ability to extract rich features and perform well on low-resolution histopathology images. Thus, our models have the potential to assist pathologists, reduce their workload, and enhance patient survival rates.

show abstract

“…Deep learning can be employed for the diagnosis of disease and its subsequent classification based on medical images [11][12][13][14][15]. Deep learning requires a lot of labeled data; however, the size of such medical image datasets is generally limited.…”

Section: Introductionmentioning

confidence: 99%

Digital holographic technique based breast cancer detection using transfer learning method

Thomas

Sheeja

2023

Journal of Biophotonics

View full text Add to dashboard Cite

The digital holographic technique is an interferometric method that provides comprehensive information on morphological traits such as cell layer thickness and shape as well as access to biophysical attributes of cells like refractive index, dry mass, and volume. This method helps characterize sample structures in three dimensions both statically and dynamically, even for transparent objects like living biological cells. This research work captures the digital holograms of breast tissues and analyzes the malignancy of the tissue using a deep learning technique. It enables dynamic measurement of the sample under investigation. Different transfer learning models such as Inception, DenseNet, SqueezeNet, VGG, and ResNet are incorporated in this work. The parameters accuracy, precision, sensitivity, and F1 score of different models are compared and found that the ResNet model outperforms better compared to other models.

show abstract

Classification of Breast Cancer using Histology images: Handcrafted and Pre-Trained Features Based Approach

Cited by 10 publications

References 13 publications

Impact of H&E Stain Normalization on Deep Learning Models in Cancer Image Classification: Performance, Complexity, and Trade-Offs

Impact of H&E Stain Normalization on Deep Learning Models in Cancer Image Classification: Performance, Complexity, and Trade-Offs

Histopathological Cancer Detection Using Intra-Domain Transfer Learning and Ensemble Learning

Digital holographic technique based breast cancer detection using transfer learning method

Contact Info

Product

Resources

About