A two-layer Conditional Random Field for the classification of partially occluded objects
This addresses the challenge of object classification under occlusion for computer vision applications, though it appears incremental as it builds on existing CRF techniques.
The paper tackles the problem of classifying partially occluded objects in images by proposing a novel two-layer Conditional Random Field framework, achieving competitive results on aerial and urban street-view datasets compared to other methods.
Conditional Random Fields (CRF) are among the most popular techniques for image labelling because of their flexibility in modelling dependencies between the labels and the image features. This paper proposes a novel CRF-framework for image labeling problems which is capable to classify partially occluded objects. Our approach is evaluated on aerial near-vertical images as well as on urban street-view images and compared with another methods.