An official website of the United States government
With the spread of COVID-19, significantly more patients have required medical diagnosis to determine whether they are a carrier of the virus. COVID-19 can lead to the development of pneumonia in the lungs, which can be captured in X-Ray and CT scans of the patient's chest. The abundance of X-Ray and CT image data available can be used to develop a high-performing computer vision model able to identify and classify instances of pneumonia present in medical scans. Predictions made by these deep learning models can increase the confidence of diagnoses made by analyzing minute features present in scans exhibiting COVID-19 pneumonia, often unnoticeable to the human eye. Furthermore, rather than teaching clinicians about the mathematics behind deep learning and heat maps, we introduce novel methods of explainable artificial intelligence (XAI) with the goal to annotate instances of pneumonia in medical scans exactly as radiologists do to inform other radiologists, clinicians, and interns about patterns and findings. This project explores methods to train and optimize state-of-the-art deep learning models on COVID-19 pneumonia medical scans and apply explainability algorithms to generate annotated explanations of model predictions that are useful to clinicians and radiologists in analyzing these images.
more »
« less
Differences between human and machine perception in medical diagnosis
Abstract Deep neural networks (DNNs) show promise in image-based medical diagnosis, but cannot be fully trusted since they can fail for reasons unrelated to underlying pathology. Humans are less likely to make such superficial mistakes, since they use features that are grounded on medical science. It is therefore important to know whether DNNs use different features than humans. Towards this end, we propose a framework for comparing human and machine perception in medical diagnosis. We frame the comparison in terms of perturbation robustness, and mitigate Simpson’s paradox by performing a subgroup analysis. The framework is demonstrated with a case study in breast cancer screening, where we separately analyze microcalcifications and soft tissue lesions. While it is inconclusive whether humans and DNNs use different features to detect microcalcifications, we find that for soft tissue lesions, DNNs rely on high frequency components ignored by radiologists. Moreover, these features are located outside of the region of the images found most suspicious by radiologists. This difference between humans and machines was only visible through subgroup analysis, which highlights the importance of incorporating medical domain knowledge into the comparison.
more »
« less
- Award ID(s):
- 1922658
- PAR ID:
- 10366492
- Author(s) / Creator(s):
- ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; more »
- Publisher / Repository:
- Nature Publishing Group
- Date Published:
- Journal Name:
- Scientific Reports
- Volume:
- 12
- Issue:
- 1
- ISSN:
- 2045-2322
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Deep neural networks (DNNs) are increasingly used in critical applications like autonomous vehicles and medical diagnosis, where accuracy and reliability are crucial. However, debugging DNNs is challenging and expensive, often leading to unpredictable behavior and performance issues. Identifying and diagnosing bugs in DNNs is difficult due to complex and obscure failure symptoms, which are data-driven and compute-intensive. To address this, we propose TransBug a framework that combines transformer models for feature extraction with deep learning models for classification to detect and diagnose bugs in DNNs. We employ a pre-trained transformer model, which has been trained in programming languages, to extract semantic features from both faulty and correct DNN models. We then use these extracted features in a separate deep-learning model to determine whether the code contains bugs. If a bug is detected, the model further classifies the type of bug. By leveraging the powerful feature extraction capabilities of transformers, we capture relevant characteristics from the code, which are then used by a deep learning model to identify and classify various types of bugs. This combination of transformer-based feature extraction and deep learning classification allows our method to accurately link bug symptoms to their causes, enabling developers to take targeted corrective actions. Empirical results show that the TransBug shows an accuracy of 81% for binary classification and 91% for classifying bug types.more » « less
-
Reflectance confocal microscopy (RCM) is a noninvasive optical imaging technique that uses a laser to capture cellular-level resolution images based on differing refractive indices of tissue elements. RCM image interpretation is challenging and requires training to interpret and correlate the grayscale output images that lack nuclear features with tissue pathology. Here, we utilize a deep learning-based framework that uses a convolutional neural network to transform grayscale images into virtually-stained hematoxylin and eosin (H&E)-like images enabling the visualization of various skin layers. To train the deep-learning framework, a series of a minimum of 7 time-lapsed, successive “stacks” of RCM images of excised tissue, spaced 1.52μm apart to a depth of 60.96μm were obtained using the Vivascope 1500. The tissue samples were stained with a 50% acetic acid solution to enhance cell nuclei. These images served as the “ground truth” to train a deep convolutional neural network with a conditional generative adversarial network (GAN)-based machine learning algorithm to digitally convert the images into GAN-based H&E-stained digital images. The machine learning algorithm was initially trained and subsequently retrained with new samples, specifically focusing on squamous neoplasms. The trained algorithm was applied to skin lesions that had a clinical differential diagnosis of squamous neoplasms including squamous cell carcinoma, actinic keratosis, seborrheic keratosis, and basal cell carcinoma. Through continuous training and refinement, the algorithm was able to produce high-resolution, histological quality images of different squamous neoplasms. This algorithm may be used in the future to facilitate earlier diagnosis of cutaneous neoplasms and enable greater uptake of noninvasive imaging technology within the medical community.more » « less
-
Abstract BackgroundLung cancer is the deadliest and second most common cancer in the United States due to the lack of symptoms for early diagnosis. Pulmonary nodules are small abnormal regions that can be potentially correlated to the occurrence of lung cancer. Early detection of these nodules is critical because it can significantly improve the patient's survival rates. Thoracic thin‐sliced computed tomography (CT) scanning has emerged as a widely used method for diagnosing and prognosis lung abnormalities. PurposeThe standard clinical workflow of detecting pulmonary nodules relies on radiologists to analyze CT images to assess the risk factors of cancerous nodules. However, this approach can be error‐prone due to the various nodule formation causes, such as pollutants and infections. Deep learning (DL) algorithms have recently demonstrated remarkable success in medical image classification and segmentation. As an ever more important assistant to radiologists in nodule detection, it is imperative ensure the DL algorithm and radiologist to better understand the decisions from each other. This study aims to develop a framework integrating explainable AI methods to achieve accurate pulmonary nodule detection. MethodsA robust and explainable detection (RXD) framework is proposed, focusing on reducing false positives in pulmonary nodule detection. Its implementation is based on an explanation supervision method, which uses nodule contours of radiologists as supervision signals to force the model to learn nodule morphologies, enabling improved learning ability on small dataset, and enable small dataset learning ability. In addition, two imputation methods are applied to the nodule region annotations to reduce the noise within human annotations and allow the model to have robust attributions that meet human expectations. The 480, 265, and 265 CT image sets from the public Lung Image Database Consortium and Image Database Resource Initiative (LIDC‐IDRI) dataset are used for training, validation, and testing. ResultsUsing only 10, 30, 50, and 100 training samples sequentially, our method constantly improves the classification performance and explanation quality of baseline in terms of Area Under the Curve (AUC) and Intersection over Union (IoU). In particular, our framework with a learnable imputation kernel improves IoU from baseline by 24.0% to 80.0%. A pre‐defined Gaussian imputation kernel achieves an even greater improvement, from 38.4% to 118.8% from baseline. Compared to the baseline trained on 100 samples, our method shows less drop in AUC when trained on fewer samples. A comprehensive comparison of interpretability shows that our method aligns better with expert opinions. ConclusionsA pulmonary nodule detection framework was demonstrated using public thoracic CT image datasets. The framework integrates the robust explanation supervision (RES) technique to ensure the performance of nodule classification and morphology. The method can reduce the workload of radiologists and enable them to focus on the diagnosis and prognosis of the potential cancerous pulmonary nodules at the early stage to improve the outcomes for lung cancer patients.more » « less
-
null (Ed.)Abstract Deep neural networks (DNNs) have achieved state-of-the-art performance in many important domains, including medical diagnosis, security, and autonomous driving. In domains where safety is highly critical, an erroneous decision can result in serious consequences. While a perfect prediction accuracy is not always achievable, recent work on Bayesian deep networks shows that it is possible to know when DNNs are more likely to make mistakes. Knowing what DNNs do not know is desirable to increase the safety of deep learning technology in sensitive applications; Bayesian neural networks attempt to address this challenge. Traditional approaches are computationally intractable and do not scale well to large, complex neural network architectures. In this paper, we develop a theoretical framework to approximate Bayesian inference for DNNs by imposing a Bernoulli distribution on the model weights. This method called Monte Carlo DropConnect (MC-DropConnect) gives us a tool to represent the model uncertainty with little change in the overall model structure or computational cost. We extensively validate the proposed algorithm on multiple network architectures and datasets for classification and semantic segmentation tasks. We also propose new metrics to quantify uncertainty estimates. This enables an objective comparison between MC-DropConnect and prior approaches. Our empirical results demonstrate that the proposed framework yields significant improvement in both prediction accuracy and uncertainty estimation quality compared to the state of the art.more » « less
