William J. Sehnert

Motion blur detection in radiographs

Image blur introduced by patient motion is one of the most frequently cited reasons for image rejection in radiographic diagnostic imaging. The goal of the present work is to provide an automated method for the detection of anatomical motion blur in digital radiographic images to help improve image quality and facilitate workflow in the radiology department. To achieve this goal, the method first reorients the image to a predetermined hanging protocol. Then it locates the primary anatomy in the radiograph and extracts the most indicative region for motion blur, i.e., the region of interest (ROI). The third step computes a set of motion-sensitive features from the extracted ROI. Finally, the extracted features are evaluated by using a classifier that has been trained to detect motion blur. Preliminary experiments show promising results with 86% detection sensitivity, 72% specificity, and an overall accuracy of 76%.

Investigation of the Variability in the Assessment of Digital Chest X-ray Image Quality

A large database of digital chest radiographs was developed over a 14-month period. Ten radiographic technologists and five radiologists independently evaluated a stratified subset of images from the database for quality deficiencies and decided whether each image should be rejected. The evaluation results showed that the radiographic technologists and radiologists agreed only moderately in their assessments. When compared against each other, radiologist and technologist reader groups were found to have even less agreement than the inter-reader agreement within each group. Radiologists were found to be more accepting of limited-quality studies than technologists. Evidence from the study suggests that the technologists weighted their reject decisions more heavily on objective technical attributes, while the radiologists weighted their decisions more heavily on diagnostic interpretability relative to the image indication. A suite of reject-detection algorithms was independently run on the images in the database. The algorithms detected 4 % of postero-anterior chest exams that were accepted by the technologist who originally captured the image but which would have been rejected by the technologist peer group. When algorithm results were made available to the technologists during the study, there was no improvement in inter-reader agreement in deciding whether to reject an image. The algorithm results do, however, provide new quality information that could be captured within a site-wide, reject-tracking database and leveraged as part of a site-wide QA program.

Adversarial Sparse-View CBCT Artifact Reduction

We present an effective post-processing method to reduce the artifacts from sparsely reconstructed cone-beam CT (CBCT) images. The proposed method is based on the state-of-the-art, image-to-image generative models with a perceptual loss as regulation. Unlike the traditional CT artifact-reduction approaches, our method is trained in an adversarial fashion that yields more perceptually realistic outputs while preserving the anatomical structures. To address the streak artifacts that are inherently local and appear across various scales, we further propose a novel discriminator architecture based on feature pyramid networks and a differentially modulated focus map to induce the adversarial training. Our experimental results show that the proposed method can greatly correct the cone-beam artifacts from clinical CBCT images reconstructed using 1/3 projections, and outperforms strong baseline methods both quantitatively and qualitatively.