We present a new approach for tracking roads from satellite images, and thereby illustrate a general computational strategy ("active testing") for tracking 1D structures and other recognition tasks in computer vision. Our approach is related to recent work in active vision on "where to look next" and motivated by the "divide-and-conquer" strategy of parlor games such as "Twenty Questions." We choose "tests" (matched filters for short road segments) one at a time in order to remove as much uncertainty as possible about the "true hypothesis" (road position) given the results of the previous tests. The tests are chosen on-line based on a statistical model for the joint distribution of tests and hypotheses. The problem of minimizing uncertainty (measured by entropy) is formulated in simple and explicit analytical terms. To execute this entropy testing rule we then alternate between data collection and optimization: at each iteration new image data are examined and a new entropy minimizat...

Abstract. Reliably extracting information from aerial imagery is a difficult problem with many practical applications. One specific case of this problem is the task of automatically detecting roads. This task is a difficult vision problem because of occlusions, shadows, and a wide variety of non-road objects. Despite 30 years of work on automatic road detection, no automatic or semi-automatic road detection system is currently on the market and no published method has been shown to work reliably on large datasets of urban imagery. We propose detecting roads using a neural network with millions of trainable weights which looks at a much larger context than was used in previous attempts at learning the task. The network is trained on massive amounts of data using a consumer GPU. We demonstrate that predictive performance can be substantially improved by initializing the feature detectors using recently developed unsupervised learning methods as well as by taking advantage of the local spatial coherence of the output labels. We show that our method works reliably on two challenging urban datasets that are an order of magnitude larger than what was used to evaluate previous approaches. 1

Recent research in perinatal pathology argues that analyzing properties of the placenta may reveal important information on how certain diseases progress. One important property is the structure of the placental blood vessels, which supply a fetus with all of its oxygen and nutrition. An essential step in the analysis of the vascular network pattern is the extraction of the blood vessels, which has only been done manually through a costly and time-consuming process. There is no existing method to automatically detect placental blood vessels; in addition, the large variation in the shape, color, and texture of the placenta makes it difficult to apply standard edge-detection algorithms. We describe a method to automatically detect and extract blood vessels from a given image by using image processing techniques and neural networks. We evaluate several local features for every pixel, such as intensity, gradient, and variance, in addition to a novel modification to an existing road detector. Pixels belonging to blood vessel regions have recognizable responses; hence, we use an artificial neural network to identify the pattern of blood vessels. A set of images where blood vessels are manually highlighted is used to train the network. We then apply the neural network to recognize blood vessels in new images. The network is effective in capturing the most prominent vascular structures of the placenta. 1