Addressing class imbalance in deep learning for small lesion detection on medical images

Bria, Alessandro; Marrocco, Claudio; Tortorella, Francesco

doi:10.1016/j.compbiomed.2020.103735

Cited by 100 publications

(61 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Sometimes, the datasets are also imbalanced, with just a small number of samples showing a particular but important condition. Bria et al [ 30 ] address the problem of class imbalance in medical images. A common technique is to use data augmentation, adding copies of some images with a transformation such as mirroring or rotation [ 31 ].…”

Section: Materials and Methodsmentioning

confidence: 99%

Study on Data Partition for Delimitation of Masses in Mammography

2021

View full text Add to dashboard Cite

Mammography is the primary medical imaging method used for routine screening and early detection of breast cancer in women. However, the process of manually inspecting, detecting, and delimiting the tumoral massess in 2D images is a very time-consuming task, subject to human errors due to fatigue. Therefore, integrated computer-aided detection systems have been proposed, based on modern computer vision and machine learning methods. In the present work, mammogram images from the publicly available Inbreast dataset are first converted to pseudo-color and then used to train and test a Mask R-CNN deep neural network. The most common approach is to start with a dataset and split the images into train and test set randomly. However, since there are often two or more images of the same case in the dataset, the way the dataset is split may have an impact on the results. Our experiments show that random partition of the data can produce unreliable training, so the dataset must be split using case-wise partition for more stable results. In experimental results, the method achieves an average true positive rate of 0.936 with 0.063 standard deviation using random partition and 0.908 with 0.002 standard deviation using case-wise partition, showing that case-wise partition must be used for more reliable results.

show abstract

Section: Materials and Methodsmentioning

confidence: 99%

Study on Data Partition for Delimitation of Masses in Mammography

2021

View full text Add to dashboard Cite

show abstract

“…Deep learning (DL) [42][43][44][45][46][47][48][49][50][51][52][53][54][55]is a class of powerful MLs based on multiple deep layers of neural networks, characterized by hundreds of layers, each of which learns to detect different features of increasing complexity from an image. In contrast to ML, DL doesn't need to define a priori a set of hand-crafted features, instead constructing its own internal features which are able to describe rich and comprehensive information, thus performing data representation and prediction jointly.…”

Section: Machine Learning and Deep Learning In Imagingmentioning

confidence: 99%

“…Microcalcifications, potentially an early sign of breast cancer, can be detected from mammograms using ML clustering methods such as kmeans [31,34] or DL, which also allows segmentation of microcalcifications [43,67] and breast parenchyma [66].…”

Section: Cancer Diagnosis and Characterizationmentioning

confidence: 99%

Artificial intelligence applications in medical imaging: A review of the medical physics research in Italy

Avanzo

Porzio²,

Lorenzon

et al. 2021

Physica Medica

View full text Add to dashboard Cite

To perform a systematic review on the research on the application of artificial intelligence (AI) to imaging published in Italy and identify its fields of application, methods and results. Materials and Methods: A Pubmed search was conducted using terms Artificial Intelligence, Machine Learning, Deep learning, imaging, and Italy as affiliation, excluding reviews and papers outside time interval 2015-2020. In a second phase, participants of the working group AI4MP on Artificial Intelligence of the Italian Association of Physics in Medicine (AIFM) searched for papers on AI in imaging. Results: The Pubmed search produced 794 results. 168 studies were selected, of which 122 were from Pubmed search and 46 from the working group. The most used imaging modality was MRI (44%) followed by CT(12%) ad radiography/mammography (11%). The most common clinical indication were neurological diseases (29%) and diagnosis of cancer (25%). Classification was the most common task for AI (57%) followed by segmentation (16%). 65% of studies used machine learning and 35% used deep learning. We observed a rapid increase of research in Italy on artificial intelligence in the last 5 years, peaking at 155% from 2018 to 2019.

show abstract

“…The classification accuracy is given by the number above each box plot pair dermoscopic images of skin lesions. The dataset is also heavily imbalanced, a common problem among medical datasets [5,12,46]. About 67% of all images belong to ''nevi'' class, while only 1% of images are instances of ''dermatofibroma''.…”

Section: Use Case: Skin Lesion Classificationmentioning

confidence: 99%

Leveraging the Bhattacharyya coefficient for uncertainty quantification in deep neural networks

Molle

Verbelen

Vankeirsbilck

et al. 2021

Neural Comput & Applic

View full text Add to dashboard Cite

Modern deep learning models achieve state-of-the-art results for many tasks in computer vision, such as image classification and segmentation. However, its adoption into high-risk applications, e.g. automated medical diagnosis systems, happens at a slow pace. One of the main reasons for this is that regular neural networks do not capture uncertainty. To assess uncertainty in classification, several techniques have been proposed casting neural network approaches in a Bayesian setting. Amongst these techniques, Monte Carlo dropout is by far the most popular. This particular technique estimates the moments of the output distribution through sampling with different dropout masks. The output uncertainty of a neural network is then approximated as the sample variance. In this paper, we highlight the limitations of such a variance-based uncertainty metric and propose an novel approach. Our approach is based on the overlap between output distributions of different classes. We show that our technique leads to a better approximation of the inter-class output confusion. We illustrate the advantages of our method using benchmark datasets. In addition, we apply our metric to skin lesion classification—a real-world use case—and show that this yields promising results.

show abstract

Addressing class imbalance in deep learning for small lesion detection on medical images

Cited by 100 publications

References 57 publications

Study on Data Partition for Delimitation of Masses in Mammography

Study on Data Partition for Delimitation of Masses in Mammography

Artificial intelligence applications in medical imaging: A review of the medical physics research in Italy

Leveraging the Bhattacharyya coefficient for uncertainty quantification in deep neural networks

Contact Info

Product

Resources

About