General Information

Ultimate Opening


In a first part, we have shown ultimate opening benefits as a generic operator of segmentation. Then we proposed several improvements making it more robust to masking and leakage problems. The performance of these improvements is illustrated in our study framework, scene-text detection and cell segmentation.

Facade Division



The second part is focused on facade analysis. Facade modeling is carried throughout a building scope. However, during image acquisition, several buildings appear in the same image. We propose an automatic method to separate different facades contained in the image. Then, we focus on semantic extraction from the facade. It consists in segment by floors, windows, balconies... to provide a realistic facede model.

Point Cloud Segmentation

The third part is focused on point cloud analysis. In urban modeling, it is necessary to introduce some elements as street furniture and pavement. We present tools for detection and classification of artifacts. These tools allow: 1- filtering out of data in order to facilitate modeling process and 2- Re-introduction of some elements (lampposts, sign boards, bus stop, etc), improving visual realism in modeled scene. We also propose an automatic method for pavement segmentation