Machine Learning for Image Analysis


Machine vision researchers have developed many image understanding systems, but their software often relies on heuristic knowledge that transfers poorly to new images, much less to new domains, and there remains considerable room for improving their behavior. One path to such improvement uses machine learning to refine or replace the handcrafted knowledge that currently guides an image understanding system.

This project aims to embed learning algorithms into an existing system that analyzes satellite images of military sites. The enhanced system incorporates an visual interface that lets the developer interactively label candidates about objects in the image, which it then converts to training data for use in learning. An important feature of the induction process is that it takes into acount the relative costs for different types of errors. Preliminary tests show the addition of machine learning improves the accuracy of building recognition over that obtained with handcrafted knowledge. Future work will include applying this approach to the recognition of lower-level components of buildings, such as U contours and junctions.

This work was funded by the Office of Naval Research through Grant N00014-94-1-0543.


Contributors to the Project

  • Dr. Kamal Ali

  • Professor Thomas Binford

  • Professor Pat Langley

  • Professor Marcus Maloof

  • Stephanie Sage

  • Related Papers

    Maloof, M. A., Langley, P., Binford, T. O., Nevatia, R., & Sage, S. (2003). Improved rooftop detection in aerial images with machine learning. Machine Learning, 53, 157-191.

    Ali, K. M., Langley, P., Maloof, M. A., Binford, T. O., & Sage, S. (1998). Improving rooftop detection with interactive visual learning. Proceedings of the Image Understanding Workshop. Monterrey, CA: Morgan Kaufmann.

    Maloof, M. A., Langley, P., Binford, T. O., & Nevatia, R. (1998). Generalizing over aspect and location for rooftop detection. Proceedings of the Fourth IEEE Workshop on Applications of Computer Vision. Princeton, NJ: IEEE Press.

    Maloof, M. A., Langley, P., Sage, S., & Binford, T. (1997). Learning to detect rooftops in aerial images. Proceedings of the 1997 Image Understanding Workshop. New Orleans: Morgan Kaufmann.

    Provan, G., Langley, P., & Binford, T.O. (1996). Probabilistic learning of three-dimensional object models. Proceedings of the Image Understanding Workshop (pp. 1403-1413). Palm Springs, CA: Morgan Kaufmann.

    Langley, P., Binford, T. O., & Levitt, T. S. (1994). Learning object models from visual observation and background knowledge. Proceedings of the Image Understanding Workshop. Monterrey, CA.

    Bowyer, K. W., Hall, L. O., Langley, P., Bhanu, B., & Draper, B. A. (1994). Report of the AAAI fall symposium on machine learning and computer vision: What, why and how. Proceedings of the Image Understanding Workshop. Monterrey, CA.

    For more information, please send email to maloof@apres.stanford.edu .