Thanks for your work.
Regarding the section 4.1 in your paper that mentions, "We adopted a similar approach here by selecting a high-quality subset of Objaverse-XL that contains 1.3 million objects. Selection is done by defining proxy estimation of human preference based on heuristics including vertex count, face count, popularity on the source website, and source of data, among other metrics," I would like to know more about your specific processing procedures or criteria. Will your processing code or criteria be released?
Thanks for your work. Regarding the section 4.1 in your paper that mentions, "We adopted a similar approach here by selecting a high-quality subset of Objaverse-XL that contains 1.3 million objects. Selection is done by defining proxy estimation of human preference based on heuristics including vertex count, face count, popularity on the source website, and source of data, among other metrics," I would like to know more about your specific processing procedures or criteria. Will your processing code or criteria be released?