Showing papers by "Veronika Cheplygina published in 2015"

Label Stability in Multiple Instance Learning

Abstract: We address the problem of instance label stability in multiple instance learning MIL classifiers. These classifiers are trained only on globally annotated images bags, but often can provide fine-grained annotations for image pixels or patches instances. This is interesting for computer aided diagnosis CAD and other medical image analysis tasks for which only a coarse labeling is provided. Unfortunately, the instance labels may be unstable. This means that a slight change in training data could potentially lead to abnormalities being detected in different parts of the image, which is undesirable from a CAD point of view. Despite MIL gaining popularity in the CAD literature, this issue has not yet been addressed. We investigate the stability of instance labels provided by several MIL classifiers on 5 different datasets, of which 3 are medical image datasets breast histopathology, diabetic retinopathy and computed tomography lung images. We propose an unsupervised measure to evaluate instance stability, and demonstrate that a performance-stability trade-off can be made when comparing MIL classifiers.

Characterizing Multiple Instance Datasets

Abstract: In many pattern recognition problems, a single feature vector is not sufficient to describe an object. In multiple instance learning (MIL), objects are represented by sets (bags) of feature vectors (instances). This requires an adaptation of standard supervised classifiers in order to train and evaluate on these bags of instances. Like for supervised classification, several benchmark datasets and numerous classifiers are available for MIL. When performing a comparison of different MIL classifiers, it is important to understand the differences of the datasets, used in the comparison. Seemingly different (based on factors such as dimensionality) datasets may elicit very similar behaviour in classifiers, and vice versa. This has implications for what kind of conclusions may be drawn from the comparison results. We aim to give an overview of the variability of available benchmark datasets and some popular MIL classifiers. We use a dataset dissimilarity measure, based on the differences between the ROC-curves obtained by different classifiers, and embed this dataset dissimilarity matrix into a low-dimensional space. Our results show that conceptually similar datasets can behave very differently. We therefore recommend examining such dataset characteristics when making comparisons between existing and new MIL classifiers. Data and other resources are available at http://www.miproblems.org.

Dissimilarity Representations for Low-Resolution Face Recognition

Abstract: Low-resolution face recognition is a very difficult problem. In this setup, the training database or gallery contains high-resolution images, but the image to be recognized is of low resolution. Thus we are dealing with a resolution mismatch problem for training and test images. Standard face recognition methods fail in this setting, which suggests that current feature representation approaches are not adequate to cope with this problem. Therefore, we propose the use of dissimilarity representations based on different strategies, which differ in how images with different resolutions are compared, to solve the resolution mismatch problem. Experiments on four standard face datasets demonstrate that a strategy based on first down-scaling and afterwards up-scaling training images while up-scaling test images outperforms all the other approaches.