220 research outputs found
Saliency Detection by Multiple-Instance Learning
No full textSaliency detection has been a hot topic in recent years. Its popularity is mainly because of its theoretical meaning for explaining human attention and applicable aims in segmentation, recognition, etc. Nevertheless, traditional algorithms are mostly based on unsupervised techniques, which have limited learning ability. The obtained saliency map is also inconsistent with many properties of human behavior. In order to overcome the challenges of inability and inconsistency, this paper presents a framework based on multiple-instance learning. Low-, mid-, and high-level features are incorporated in the detection procedure, and the learning ability enables it robust to noise. Experiments on a data set containing 1000 images demonstrate the effectiveness of the proposed framework. Its applicability is shown in the context of a seam carving application
Automated Intervertebral Disc Detection from Low Resolution, Sparse MRI Images for the Planning of Scan Geometries
No full textRobust and accurate identification of intervertebral discs from low resolution, sparse MRI scans is essential for the automated scan planning of the MRI spine scan. This paper presents a graphical model based solution for the detection of both the positions and orientations of intervertebral discs from low resolution, sparse MRI scans. Compared with the existing graphical model based methods, the proposed method does not need a training process using training data and it also has the capability to automatically determine the number of vertebrae visible in the image. Experiments on 25 low resolution, sparse spine MRI data sets verified its performance
Tracking vehicles as groups in airborne videos
No full textAirborne vehicle tracking system is receiving increasing attention due to its high mobility, low cost and large surveillance scope. However, tracking multiple vehicles simultaneously on airborne platform is a challenging problem, owing to camera vibration, which causes visible frame-to-frame jitter in the airborne videos and uncertain vehicle motion. To address these problems, a new collaborative tracking framework is proposed in this paper. The framework consists of a two-level tracking process to track vehicles as groups. The higher level builds the relevance network and divides target vehicles into different groups, where the relevance is calculated based on the status information of vehicles obtained from the lower level. The proposed group tracking takes into account the relevance between vehicles and reduces the impact of camera vibration. Experimental results demonstrated that the proposed method has better performance in terms of tracking speed and tracking accuracy compared to other existing approaches based on particle filter and stationary grouping
Global structure constrained local shape prior estimation for medical image segmentation
No full textOrgan shape plays an important role in clinical diagnosis, surgical planning and treatment evaluation. Shape modeling is a critical factor affecting the performance of deformable model based segmentation methods for organ shape extraction. In most existing works, shape modeling is completed in the original shape space, with the presence of outliers. In addition, the specificity of the patient was not taken into account. This paper proposes a novel target-oriented shape prior model to deal with these two problems in a unified framework. The proposed method measures the intrinsic similarity between the target shape and the training shapes on an embedded manifold by manifold learning techniques. With this approach, shapes in the training set can be selected according to their intrinsic similarity to the target image. With more accurate shape guidance, an optimized search is performed by a deformable model to minimize an energy functional for image segmentation, which is efficiently achieved by using dynamic programming. Our method has been validated on 2D prostate localization and 3D prostate segmentation in MRI scans. Compared to other existing methods, our proposed method exhibits better performance in both studies. (C) 2013 Elsevier Inc. All rights reserved
A bridged denoising convolutional neural network (BD-CNN) for photon-counting CT
No full textMay 2019School of EngineeringTraditional monochromatic CT scanners have served as one of the most critical medical diagnostic instruments. Recently, the development of photon-counting CT brings spectral information to the otherwise black-and-white CT images, allowing the separation of materials. However, noise from a range of sources still degrades the overall image quality. Inspired by the successes of deep learning implementations in image processing, a bridged denoising convolutional neural network (BD-CNN) is proposed, aiming to reduce the noise presented in the image and enhance the overall image quality quantified by PSNR, RMSE, and SSIM. The BD-CNN is trained with a simulated labeled dataset of photon-counting CT. The results demonstrate that the denoising network improves the overall quality of the images and the noise presented in the original noisy images is decreased significantly. Compared to iterative denoising algorithms (Median filtering, BM3D), the proposed method not only achieved better results measured by the metrics, but it also demonstrates better consistency in image denoising. In conclusion, the proposed BD-CNN is a promising deep learning architecture for image denoising and can be further improved with minor adjustments. Further studies are needed for a more suitable loss function for deep learning in medical imaging and different image quality metrics optimized for comparing diagnostic value of medical images.M
Privacy-Preserving Federated Brain Tumour Segmentation
No full textDue to medical data privacy regulations, it is often infeasible to collect and share patient data in a centralised data lake. This poses challenges for training machine learning algorithms, such as deep convolutional networks, which often require large numbers of diverse training examples. Federated learning sidesteps this difficulty by bringing code to the patient data owners and only sharing intermediate model training updates among them. Although a high-accuracy model could be achieved by appropriately aggregating these model updates, the model shared could indirectly leak the local training examples. In this paper, we investigate the feasibility of applying differential-privacy techniques to protect the patient data in a federated learning setup. We implement and evaluate practical federated learning systems for brain tumour segmentation on the BraTS dataset. The experimental results show that there is a trade-off between model performance and privacy protection costs.</p
Image-driven fact-checking of ai generated chest radiology reports
No full textAugust2025School of EngineeringWith the developments in radiology artificial intelligence (AI), many researchers have turnedto the problem of automated reporting of imaging studies. The goal of such work is to
produce a preliminary read of imaging studies in locations such as emergency rooms where
a radiologist may not be readily available, or to present a preliminary structured report
to radiologists to reduce their dictation workload. An automatically produced structured
report could also be more consistent and easier to read, leading to improved accuracy and
lower overall costs of radiology reads in clinical workflows.
Among the imaging areas where this has been found most useful are chest X-rays,
which are the most common imaging modality read by radiologists in hospitals and tele-
radiology practices today. With the recent rise of generative AI, a number of researchers and
corporations are attempting to generate preliminary reports for chest X-ray images thanks
to the availability of relatively large datasets such as MIMIC and CheXpert that come with
their companion reports for training large vision-language models (VLMs). These newly
emerged VLMs can generate longer and more natural sentences when prompted with good
radiology-specific linguistic cues. However, despite the powerful language generation capabil-
ities, ensuring there are no hallucinations, incorrect mentions of findings or their descriptions,
has been difficult for these models limiting their clinical applicability. While methods for
hallucination removal and fact-checking exist for large language models, with strategies such
as direct policy optimization (DPO) or proximal policy optimization (PPO), and reward
models, they are mostly applicable during training or fine-tuning of the models. On the
other hand, methods that check facts during inference time often consult external general
knowledge or detect errors through analysis of produced text either by themselves or through
an LLM serving as a judge. In radiology report generation, however, neither is possible since
the report has to be specific to the patient and consistent with the evidence seen in the
imaging. Since the automated reporting LLMs themselves have hallucinations, there are no
teacher LLMs that are good enough to correct automatically generated radiology reports.
Further, they may not be able to corroborate their deductions with the patient-specific im-
age. Finally, any fact checking should be agnostic to the radiology report generation tool
to give versatility of use during clinical deployment where different choices of vendors may be prevalent with separate evolving capabilities over time. Thus, there is a need to develop
an independent fact-checking method for use during clinical inference to bootstrap radiology
report generation and increase their adoption in clinical workflows.
This Master’s thesis investigates a hypothesis that it is possible to develop such inde-
pendent discriminative neural networks as fact-checking models for use during inference to
detect and correct errors in automatically generated reports. The key idea explored in the
thesis is that by creating a synthetic dataset of real and fake findings derived from ground
truth reports and pairing them with the corresponding chest X-ray images, a fact-checking
classifier could be trained to distinguish between real/correct description of findings and
incorrect description of findings when they are paired with the corresponding images. Such
an independently developed classifier can then be used to detect and correct errors in the
reports generated by automated radiology reporting tools.
To proceed with the verification of the hypothesis, the thesis is divided into 4 investi-
gations. First, by examining several radiology reporting methods, we analyze the types of
errors made by the report generators to conclude four major error types such as irrelevant
predictions, polarity reversal or omissions, incorrect location predictions and other types such
as incorrect severity assessments. We then simulate the errors to create a large synthetic
dataset by perturbing findings and their locations in ground truth reports reflecting real and
fake findings-location pairs with images. We then proceed to build a discriminative classifier
to detect the errors and remove the finding errors in reports using two different methods,
one that is based on the findings alone and the other that captures their spatial locations.
Finally, we develop methods to correct the automated report while still ensuring language
correctness by careful prompting of a large language model using information derived from
the fact checking model.
Throughout, we conduct experiments with multiple benchmark datasets and conduct
ablation experiments to select relevant architectural configurations and document the overall
improvement in the quality of the report by the use of our fact-checking model to detect
and correct errors. A novel measure was developed for assessing the report correctness
leveraging both clinical accuracy and phrase grounding accuracy. Explainable visualizations
were generated to show the deviation of the reported findings from predicted findings and
their locations generated by the fact-checking model.
The overall results indicated that it was possible to develop a fact-checking model using an independently collected dataset of real and fake findings to simulate the errors made by
report generators. The resulting fact-checking model was over 90% accurate as tested on
multiple benchmark datasets and led to improvement in the quality of the automatically
generated reports in the range of 7-29%. A high degree of concordance was found between
the use of our fact-checking model and ground truth for verification of automated reports
leading us to also conclude that the fact-checking model has the potential to serve as a
surrogate ground truth during clinical inference. This proves further utility of our model
as an additional validation checkpoint in making AI models robust and ready for clinical workflows.M
Stabilizing Deep Tomographic Reconstruction Networks
No full text# Stabilizing Deep Tomographic Reconstruction Networks
# This repository contains the code, mentioned networks and test datasets from the paper "Stabilizing Deep Tomographic Reconstruction Networks" by W. Wu, et al.
# The code is divided into two modalities, i.e., CT and MRI, corresponding to two folders named by CT and MRI. ACID is a framework, the authors can use the framework based on themselves trained works.
#If you use the code, please cite our work
@article{Wu 2020,
title={ Stabilizing Deep Tomographic Reconstruction Networks },
author={ Weiwen Wu,Dianlin Hu, Wenxiang Cong,Hongming Shan,Shaoyu Wang,Chuang Niu,Pingkun Yan,Hengyong Yu,Varut Vardhanabhutiand Ge Wang },
journal={arXiv preprint arXiv: 2008.01846},
year={2020}
}
# CT folder: There are 11 sub-folder and Testmain.m. To run this code, you need to ensure your computer or work station run FBPConvNet, which can be downloaded publically from https://github.com/panakino/FBPConvNet. The lib subfolder should be added into path.
# Run Testmain.m to fast generate the reconstruction results with modifying the path. ACID subfolder contains ACID reconstruction demos for structure-changes, tiny-perturbation, more-input-data and ACID against whole Adversarial attack. Ablation subfolder is used to generate the ablation results. Demo_adversarial_pert_ACID and Demo_adversarial_pert_NN are used to adversarial attacks from the whole ACID and a single NN, where Demo_adversarial_pert_NN is sorted out based on Antun, Vegard, et al. "On instabilities of deep learning in image reconstruction and the potential costs of AI."?PNAS, 117.48 (2020): 30088-30095. Run ACIDFindPerMain.m to find the adversarial attack for whole ACID and run Demo_adversarial_pert_NN_ELL for generating the adversarial attack for Ell-50.
# CS-based and dictionary learning-based reconstruction methods are also included
# Testdata and Out_data subfolder are used to store inputdata and reconstruction results.
# Environment: Window 10 system, Matlab 2017b, Matconvnet-1.0-beta23, cuda 10.0
# MRI folder: these files focus on MRI reconstruction. There are three methods related to deep-learning-based MRI reconstruction in our paper, including AUTOMAP, DAGAN, ADMM-Net and the traditional method TGV. Their reconstruction results used in the reference are included in this folder.
# You can reproduce the results by downloading all the files and configure your workstation following the instruction of different established reconstruction methods, such as AUTOMAP, DAGAN, ADMM-Net.
# All the test data can be found in "InputData" and all the reconstruction images can be found in "ReconResult". Specified environment depending on network environment, for example, ACID building in DAGAN depends on Windows 10 system, TensorFlow 1.8.0, cuda 10.0
#If you have any problems, please contact with [email protected]; [email protected] or any one of co-authors
- …
