Learning Separable Filters.

Learning filters to produce sparse image representations in terms of over-complete dictionaries has emerged as a powerful way to create image features for many different purposes. Unfortunately, these filters are usually both numerous and non-separable, making their use computationally expensive. In this paper, we show that such filters can be computed as linear combinations of a smaller number of separable ones, thus greatly reducing the computational complexity at no cost in terms of performance. This makes filter learning approaches practical even for large images or 3D volumes, and we show that we significantly outperform state-of-the-art methods on the curvilinear structure extraction task, in terms of both accuracy and speed. Moreover, our approach is general and can be used on generic convolutional filter banks to reduce the complexity of the feature extraction step.

Supervised feature learning for curvilinear structure segmentation.

We present a novel, fully-discriminative method for curvilinear structure segmentation that simultaneously learns a classifier and the features it relies on. Our approach requires almost no parameter tuning and, in the case of 2D images, removes the requirement for hand-designed features, thus freeing the practitioner from the time-consuming tasks of parameter and feature selection. Our approach relies on the Gradient Boosting framework to learn discriminative convolutional filters in closed form at each stage, and can operate on raw image pixels as well as additional data sources, such as the output of other methods like the Optimally Oriented Flux. We will show that it outperforms state-of-the-art curvilinear segmentation methods on both 2D images and 3D image stacks.

Accurate and efficient linear structure segmentation by leveraging ad hoc features with learned filters.

Extracting linear structures, such as blood vessels or dendrites, from images is crucial in many medical imagery applications, and many handcrafted features have been proposed to solve this problem. However, such features rely on assumptions that are never entirely true. Learned features, on the other hand, can capture image characteristics difficult to define analytically, but tend to be much slower to compute than handcrafted features. We propose to complement handcrafted methods with features found using very recent Machine Learning techniques, and we show that even few filters are sufficient to efficiently leverage handcrafted features. We demonstrate our approach on the STARE, DRIVE, and BF2D datasets, and on 2D projections of neural images from the DIADEM challenge. Our proposal outperforms handcrafted methods, and pairs up with learning-only approaches at a fraction of their computational cost.