The Generic Region Merging library is now a remote module for OTB

August 2015, 18th

New Algorithm problem: The Last Shall Be the First

July 2015, 18th

New Algorithm problem: Birds of a feather flock together

July 2015, 8th

New Algorithm problem: Two numbers are ever quite alike!

June 2015, 29th

GRM Library: two new criteria available

June 2015, 9th

Two new homogeneity criteria have been added in the GRM Library:

You can download the GRM library here:

GRM Library: new stable version

June 2015, 2nd

A new version of the GRM library is available. The code has been tested and major modifications have been done:

  • A memory bug has been corrected due to circular references with the std::shared_ptr.
  • The Contour encoder seems to contain a bug, then it has been removed and will be added later. Instead each segment contains a list of its internal pixels. However, be careful since this current solution requires more memory.
  • At last but not least, the computation of the merging costs has been optimized making the algorithm really more efficient (10x speedup).
  • Warning: the LSRM library available has not been corrected and contains some bugs. This will be corrected soon.
  • Do not hesitate to send me some suggestions to improve the GRM library by adding some new functionalities. Or, even better you can commit your own extensions by sending me a message with your files. After verification, it will be added.

LSRM Library: (Beta version may contain some bugs)

May 2015, 21th

In the frame of my Ph.D., I decide to develop the Large Scale Region-Merging library. The goal of this library is to segment large satellite images which do not fit in the main memory. The library is still in development and may contain some bugs. Do not hesitate to try it for your own research and send me feedback about how we can improve it. Currently this library has been tested successfully for the segmentation of Pléiades scenes.

New Algorithm problem: palindrome number

May 2015, 21th

A new nice problem which can be solved in an elegant way using recursion.

New Publication

May 2015, 14th

The paper "A Scalable Tile-Based Framework for Region-Merging Segmentation" is now available in the IEEE Xplore Digital library in Open Access.

Click here to access to the paper

Processing large very high-resolution remote sensing images on resource-constrained devices is a challenging task because of the large size of these data sets. For applications such as environmental monitoring or natural resources management, complex algorithms have to be used to extract information from the images. The memory required to store the images and the data structures of such algorithms may be very high (hundreds of gigabytes) and therefore leads to unfeasibility on commonly available computers. Segmentation algorithms constitute an essential step for the extraction of objects of interest in a scene and will be the topic of the investigation in this paper. The objective of the present work is to adapt image segmentation algorithms for large amounts of data. To overcome the memory issue, large images are usually divided into smaller image tiles, which are processed independently. Region-merging algorithms do not cope well with image tiling since artifacts are present on the tile edges in the final result due to the incoherencies of the regions across the tiles. In this paper, we propose a scalable tile-based framework for region-merging algorithms to segment large images, while ensuring identical results, with respect to processing the whole image at once. We introduce the original concept of the stability margin for a tile. It allows ensuring identical results to those obtained if the whole image had been segmented without tiling. Finally, we discuss the benefits of this framework and demonstrate the scalability of this approach by applying it to real large images.