- Related item:
- Searching for Structures of Interest in an Ultrasound Video Sequence
- Description:
-
Ultrasound diagnosis and therapy is typically protocol driven but often criticized for requiring highly-skilled sonographers. However there is a shortage of highly trained sonographers worldwide, which is limiting the wider adoption of this cost-effective technology. The challenge therefore is to make the technology easier to use. We consider this problem in this paper. Our approach combines simple standardized clinical US scanning protocols (defined by our clinical partners) with machine learning driven image analysis solutions to enable a non-expert to perform ultrasound-based diagnostic tasks with minimal training. Motivated by recent work on dynamic texture analysis within the computer vision community, we have developed, and evaluated on clinical data, a framework that given a training set of Ultrasound Sweep Videos (USV), models the temporal evolution of objects of interest as a kernel dynamic texture which can form the basis of a metric for detecting structures of interest in new unseen videos. We describe the full original method, and demonstrate that it outperforms a simpler recently proposed approach on phantom data, and is significantly superior in performance on real clinical data.
- Related item:
- Fisher Vector Encoding for Detecting Objects of Interest in Ultrasound Videos
- Description:
-
One of the main factors limiting the wider adoption of ultrasound imaging for diagnosis and therapy is requiring highly skilled sonographers. In this paper we consider the challenge of making this technology easier to use for non-experts. Our approach follows some of the recently proposed frameworks that break the process into firstly data acquisition through a simple and task-specific scan protocol followed by using machine learning methodologies to assist non-experts in performing diagnostic tasks. We present an object classification pipeline to identify the fetal skull, heart and abdomen from all the other frames in an ultrasound video, using Fisher vector features. We describe the full proposed method and provide a comparison with a recently proposed approach based on Bag of Visual Words (BoVW) to demonstrate that the new approach is superior in terms of accuracy (98.9% versus 87.1%).
- Related item:
- Automatic fetal head detection on video clips from a low-cost portable USB ultrasound (US) device
- Description:
-
The use of ultrasound (US) is less prevalent in low income countries, in part due to lack of availability of trained staff; and because of high costs associated with US devices. Low cost US probes connected to laptop computers using dedicated software are currently available and can be used for real time fetal scans. In this paper we have investigated whether a combined machine learning and data acquisition approach to fetal head detection from a low cost USB probe is comparable to a high end probe.
- Related item:
- Towards automating the ISUOG “six-step basic ultrasound” scan
- Description:
-
Due to a global shortage of trained sonologists, it is important to develop simple but accurate scan protocols for the detection of pregnancy risks. Recent work has suggested a six-step protocol for basic fetal ultrasound (US) suitable for resource-constrained settings. Here we report progress towards automating the first two steps: detection of the fetal presentation and viability. We have approached this problem by combining the simple clinical protocol with machine learning solutions.
- Related item:
- Object Classification in an Ultrasound Video Using LP-SIFT Features
- Description:
-
The advantages of ultrasound (US) over other medical imaging modalities have provided a platform for its wide use in many medical fields, both for diagnostic and therapeutic purposes. However one of the limiting factors which has affected wide adoption of this cost-effective technology is requiring highly skilled sonographers and operators. We consider this problem in this paper which is motivated by advancements within the computer vision community. Our approach combines simple and standardized clinical ultrasound procedures with machine learning driven imaging solutions to provide users who have limited clinical experience, to perform simple diagnostic decisions (such as detection of a fetal breech presentation). We introduce LP-SIFT features constructed using the well-known SIFT features, utilizing a set of feature symmetry filters. We also illustrate how such features can be used in a bag of visual words representation on ultrasound images for classification of anatomical structures that have significant clinical implications in fetal health such as the fetal head, heart and abdomen, despite the high presence of speckle, shadows and other imaging artifacts in ultrasound images.
