Computer aided diagnosis systems are used to assist radiologists in their decision making. The sensitivity of these systems is hindered by the complexity of the structures inside the lungs. Several systems and methods have been proposed to detect and classify lung nodules, but all of them have their strengths and weaknesses. One way to overcome the weaknesses is to combine multiple systems. Systems based on handcrafted features capture a limited set of characteristics from the image, while deep learning based classifiers can deal with a wider range of structures. In this work, several ways to combine a handcrafted feature based classifier with four convolutional neural network are explored. The systems were combined merging the probabilities assigned to the detections in several ways. Support-vector machine, multilayer perceptron and random forest classifiers were used to combine the selected classifiers. The LUNA16 Challenge was used to evaluate the performance of the resulting hybrid systems. In all cases, the hybrid systems outperformed the individual systems. Although the average of sensitivities are similar for most of the combinations, the best hybrid system achieves a gain of 35 extra nodules at 4 FP per scan.
Convolutional neural networks are known to require large amounts of data to achieve optimal performance. In addition, data is commonly computationally augmented using a variety of geometric and intensity transformations to further extent the set of training samples. In medical imaging, annotated data is often scarce or costly to obtain, and there is considerable interest in methods to reduce the amount of data needed. In this work, we investigate the relative benefit of increasing the amount of original data, with respect to computationally augmenting the amount of training samples, for the case of false positive reduction of lung nodules candidates. To this end, we have implemented a previously published topology for classification, shown to achieve state of the art results on the publicly available Luna16 dataset. Numerous models were trained using different amounts of unique training samples and different degrees of data augmentation involving rotations and translations, and the performance was compared. Results indicate that in general, better performance is achieved when increasing the amount of data, or augmenting the data more extensively, as expected. Surprisingly however, we observed that after reaching a certain amount of unique training samples, data augmentation leads to significantly better performance compared to adding the same number of new samples to the training dataset. We hypothesize that the augmentation has aided in learning more general {rotation and translation invariant-features, leading to improved performance on unseen data. Future experiments include more detailed characterization of this behavior, and relating this to the topology and amount of parameters to be trained.
Recently, efficient image descriptors have shown promise for image classification tasks. Moreover, methods based on the combination of multiple image features provide better performance compared to methods based on a single feature. This work presents a simple and efficient approach for combining multiple image descriptors. We first employ a Naive-Bayes Nearest-Neighbor scheme to evaluate four widely used descriptors. For all features, “Image-to-Class” distances are directly computed without descriptor quantization. Since distances measured by different metrics can be of different nature and they may not be on the same numerical scale, a normalization step is essential to transform these distances into a common domain prior to combining them. Our experiments conducted on a challenging database indicate that z-score normalization followed by a simple sum of distances fusion technique can significantly improve the performance compared to applications in which individual features are used. It was also observed that our experimental results on the Caltech 101 dataset outperform other previous results.
Microcalcifications are tiny spots of calcium deposit that often occur in female breasts. Microcalcifications are common in healthy woman, but they often are an early sign of breast cancer. On a mammogram; the current standard of care for breast screening; calcifications appear as tiny white dots. They may occur scattered throughout the breast or grouped in clusters. Radiologists determine the suspiciousness based upon several factors, including position, frequency, grouping, evolution compared to prior studies and shape. In this paper, we study micro-CT images of biopsy samples containing microcalcifications. The scanner delivers 3D images with a voxel size of 8.66 μm, i.e. ca. 8 times the spatial resolution of a contemporary digital mammogram. We propose an automated binary classification method of the samples, based upon shape analysis of the microcalcifications. The study is performed on a set of 50 benign and 50 malign samples preserved in paraffin. The ground truth of the classification is based upon anapathological investigation of the paraffin blocks. The results show a sensitivity, i.e. the percentage of correctly classified malign samples, of up to 98% with a specificity of 40%.
This paper provides an overview of the medical scales which are currently in practice at the geriatrics department
of the hospital for assessing independence and mobility of elderly patients. Several shortcomings and issues related
to the scales are identified. It is shown how a 3D camera system could be used for the automatic assessment
of several items of the scales. this automated assessment is overcoming many of the issues with the existing
methods. An analysis of the automatically identified activity features of a typical patient is used to compare the
data derived from our system with data obtained with accelerometer readings.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.