As SPIE medicalimaging celebrates its 50th anniversary, we reflect on the history of the imageperception, observer performance, and technology assessmentconference and its importance within the SPIE medicalimaging...
详细信息
The proceedings contain 39 papers. The topics discussed include: using gradient of Lagrangian function to compute efficient channels for the ideal observer;effects of feature selection and internal-noise levels for a ...
ISBN:
(纸本)9781510685963
The proceedings contain 39 papers. The topics discussed include: using gradient of Lagrangian function to compute efficient channels for the ideal observer;effects of feature selection and internal-noise levels for a search-capable model observer;combining image texture and morphological features in low resource perception models for signal detection tasks;perceived color contrast metrics for clinical images;assessment of cell nuclei AI foundation models in kidney pathology;does concurrent reading with ai lead to more false negative errors for cancers that are not marked by AI?;and dual roles of calcification features in the Mirai mammographic breast cancer risk prediction model: early micro-calcification detection and identification of high-risk calcifications.
The proceedings contain 38 papers. The topics discussed include: recognition of radiological decision errors from eye movement during chest X-ray readings;A hybrid CNN-Swin transformer network as deep learning model o...
ISBN:
(纸本)9781510671621
The proceedings contain 38 papers. The topics discussed include: recognition of radiological decision errors from eye movement during chest X-ray readings;A hybrid CNN-Swin transformer network as deep learning model observer to predict human observer performance in 2AFC trial;adaptive learning approach to improve generalization performance of a domain-aware CNN-based ideal model observer;addition of a threshold mechanism to model observers for medicalimage quality assessment;investigation of different model observers for including signal-detectability in the training of CNNs for CT image reconstruction;computed tomography optimization using a volumetric channelized hoteling observer approach for energy integrating and photon-counting CT scanners;Sequestration of imaging studies in MIDRC: using load factor to minimize algorithm performance overestimation and image reuse;and predicting the gist of breast cancer on a screening mammogram using global radiomic features.
The proceedings contain 48 papers. The topics discussed include: comparing experts to novices: reduced satisfaction of search when searching with virtual breast tomosynthesis;optimal visual search strategy with inter-...
ISBN:
(纸本)9781510660397
The proceedings contain 48 papers. The topics discussed include: comparing experts to novices: reduced satisfaction of search when searching with virtual breast tomosynthesis;optimal visual search strategy with inter-saccade response correlations;challenges and solutions to processing and visualizing eye tracking data from digital pathology studies;identifying and preventing fatigue in digital breast tomosynthesis;global mammographic radiomic signature can predict radiologists’ difficult-to-interpret normal cases;a comparative study of diagnostic performance and work experience of radiologists in three countries interpreting digital breast tomosynthesis;false-negative diagnosis might occur due to absence of the global radiomic signature of malignancy on screening mammograms;investigating the error-making patterns in reading high-density screening mammograms between radiologists from two countries;developing and assessing an AI-based multi-task prediction system to assist radiologists detecting lung diseases in reading chest x-ray images;and Obuchowski-Rockette analysis for multi-reader multi-case (MRMC) readers-nested-in-test study design with unequal numbers of readers.
The proceedings contain 48 papers. The topics discussed include: developing interactive computer-aided detection tools to support translational clinical research;assessment of manual and automated intracranial artery ...
ISBN:
(纸本)9781510649453
The proceedings contain 48 papers. The topics discussed include: developing interactive computer-aided detection tools to support translational clinical research;assessment of manual and automated intracranial artery diameter measurements;analyzing neural networks applied to an anatomical simulation of the breast;using virtual clinical trials to determine the accuracy of ai-based quantitative imaging biomarkers in oncology trials using standard-of-care CT;interpretable deep learning models for better clinician-ai communication in clinical mammography;investigating reading strategies and eye behaviors associated with high diagnostic performance when reading digital breast tomosynthesis (DBT) images;satisfaction of search (SOS) error and new lesions identification on imaging in central review for clinical trials;and a deep Q-learning method for optimizing visual search strategies in backgrounds of dynamic noise.
The proceedings contain 30 papers. The topics discussed include: advancing the AmbientGAN for learning stochastic object models;CNN based anthropomorphic model observer for defect localization;understanding CNN based ...
ISBN:
(纸本)9781510640276
The proceedings contain 30 papers. The topics discussed include: advancing the AmbientGAN for learning stochastic object models;CNN based anthropomorphic model observer for defect localization;understanding CNN based anthropomorphic model observer using classification images;a hybrid channelized hoteling observer for estimating the ideal linear observer for total-variation-based image reconstruction;implementation of CNN-based multi-slice model observer for 3D cone beam CT;supervised learning-based ideal observer approximation for joint detection and estimation tasks;GAN generated model observer for one class detection in SPECT imaging;modeling human observer detection in undersampled magnetic resonance imaging (MRI);observer models utilizing compressed textures;and effect of CAD system with a vessel suppression function on clinical lung nodule detection in chest CT scans.
It is widely accepted that the Bayesian ideal observer (IO) should be used to guide the objective assessment and optimization of medicalimaging systems. The IO employs complete task-specific information to compute te...
详细信息
ISBN:
(纸本)9781510685963;9781510685970
It is widely accepted that the Bayesian ideal observer (IO) should be used to guide the objective assessment and optimization of medicalimaging systems. The IO employs complete task-specific information to compute test statistics for making inference decisions and performs optimally in signal detection tasks. However, the IO test statistic typically depends non-linearly on the image data and cannot be analytically determined. The ideal linear observer, known as the Hotelling observer (HO), can sometimes be used as a surrogate for the IO. However, when image data are high dimensional, HO computation can be difficult. Efficient channels that can extract task-relevant features have been investigated to reduce the dimensionality of image data to approximate IO and HO performance. This work proposes a novel method for generating efficient channels by use of the gradient of a Lagrangian-based loss function that was designed to learn the HO. The generated channels are referred to as the Lagrangian-gradient (L-grad) channels. Numerical studies are conducted that consider binary signal detection tasks involving various backgrounds and signals. It is demonstrated that channelized HO (CHO) using L-grad channels can produce significantly better signal detection performance compared to the CHO using PLS channels. Moreover, it is shown that the proposed L-grad method can achieve significantly lower computation time compared to the PLS method.
Various supervised learning-based medicalimage reconstruction methods have been developed with the goal of improving image quality (IQ). These methods typically use loss functions that minimize pixel-level difference...
详细信息
ISBN:
(纸本)9781510685963;9781510685970
Various supervised learning-based medicalimage reconstruction methods have been developed with the goal of improving image quality (IQ). These methods typically use loss functions that minimize pixel-level differences between the reconstructed and high-quality target images. While they may seemingly perform well based on traditional image quality metrics such as mean squared error, they do not consistently improve objective IQ measures based on diagnostic task performance. This work introduces a task-informed learned image reconstruction method. To establish the method, a measure of signal detection performance is incorporated in a hybrid loss function that is used for training. The proposed method is inspired by null space learning, and a task-informed data-consistent (DC) U-Net is utilized to estimate a null space component of the object that enhances task performance, while ensuring that the measurable component is stably reconstructed using a regularized pseudo-inverse operator. The impact of changing the specified task or observer at inference time to be different from that employed for model training, a phenomenon we refer to as "task-shift" or "observer-shift", respectively, was also investigated.
Texture analysis holds significant importance in various imaging fields due to its ability to provide statistical, structural, and intrinsic spatial information from images. In this work, we examine several first and ...
详细信息
ISBN:
(纸本)9781510685963;9781510685970
Texture analysis holds significant importance in various imaging fields due to its ability to provide statistical, structural, and intrinsic spatial information from images. In this work, we examine several first and second-order texture features on simulated and clinical DBT images. We examined some essential characteristics of texture features that show higher discriminatory potential for mass detection in digital breast tomosynthesis. We further examined the use of these texture features along with morphological features in a two stage visual search (VS) model observer for mass detection in DBT. Our preliminary results show that incorporation of texture features reduced the number of suspicious locations in the first stage of VS model. Our preliminary results with an eye tracking system and observer gaze points align well with the "search" regions predicted by either the texture aided or thresholded VS observer. In summary, we show how additing perceptually relevant texture features or a thresholding mechanism enhances our visual search observer models. Future work will examine feature selections for changing tasks.
This study aims to investigate how the fluctuation of time intervals between self-assessment test sets influence the performance of radiologists and radiology trainees. The data was collected from 54 radiologists and ...
详细信息
暂无评论