Domain-Specific Models in Computer Vision

摘要

The grant supported research on motion, shape, geometry, and object recognition.Work in the first area dealt with geometric and probabilistic models for understanding image sequences. In the second area, growth and migration models for the generation and evolution of shapes were developed. The third area included a wide variety of basic geometrical studies. The fourth area dealt with an integrated approach to model-based object segmentation and recognition, and with the recognition of object functionality. In addition, annual, subject-indexed bibliographies of the literature on computer vision and image analysis-about 1500 references per year-were compiled and published.