Questions for "Large-Scale Unsupervised Object Discovery"

-----------------------------------------------------------------------------------------------

Please send your answers to: galessos@cs.uni-freiburg.de by July 1 at 13:15.

1. What limitation of the related approaches prevents them from scaling up to large datasets, and how do the authors of LOD overcome it? (2 sentences)

2. What do the "quadratic formulation" and the "PageRank formulation" have in common in their fundamental way to solve the problem at hand? Based on the text, can you provide a brief intuitive explanation of why this is the case? (2-3 sentences)

3. What is, in your opinion, the experimental result that best highlights the advantages of LOD? (1-2 sentences)