elementdiscovery | Visual Element Discovery as Discriminative Mode Seeking

elementdiscovery is a C++ library. elementdiscovery has no bugs, it has no vulnerabilities, it has a Permissive License and it has low support. You can download it from GitHub.

This code is the author's implementation of the algorithm presented in:. Carl Doersch, Abhinav Gupta, and Alexei A. Efros, "Mid Level Visual Element Discovery as Discriminative Mode Seeking" in NIPS 2013. The majority of the code was written by Carl Doersch (cdoersch at cs dot cmu dot edu), although there are major contributions from Saurabh Singh and some from others who are noted in the code. This is officially unsupported research code, but our goal is that this will be as useful as possible. You are encouraged to ask questions via e-mail, and strongly encouraged to give feedback if you find the code counter-intuitive. I plan to update this code as issues are discovered. Running the code on indoor67 should require about 8GB of RAM per machine. indoor67_main will attempt to estimate the number of jobs to run on each machine based on the RAM each machine has. The program also needs about 30GB of free disk space and about 300GB in the temporary local directories (that's total; it can be distributed across different machines). If you're not using local directories, then that 300GB will need to be on the shared filesystem.