SPAN is a new and simple backbone for computational pathology. Same backbone drops into classification and survival prediction (SPAN-MIL), segmentation (SPAN-UNet), and beats SOTA. It bridges the domain gap between general vision intelligence methods and slide level analysis.
Single-scale input → multi-scale hierarchy for gigapixel pathology. We spend compute only where the tissue actually is.
#CVPR2026 Poster #211, Fri 6/6 morning,come say hi!
💻Code: https://t.co/vgV9mOgmWn
Paper: https://t.co/adl9pV5nef #DigitalPathology#ComputationalPathology
@ICCVConference Only 1/3 final ratings in so far—received just yesterday.
The other two remain unchanged despite reviewers mentioning updates post-rebuttal.
The wait is really discouraging after all the effort.
@CVPR I'm sorry for bothering you these past few days, but perhaps this is just how the world is. They ceremoniously designed guidelines, ceremoniously punished irresponsible reviewers, yet support obviously absurd irresponsible ACs. I've learned my lesson
@CVPR Received a short response from PCs. They're defending the ACs' decision even though their meta-review is wrong, ridiculous and violates CVPR's reviewers' guidelines. I feel that my expectations for fairness, as well as efforts in reviewing and writing the rebuttal, like a joke.
@CVPR Are appeals still under consideration, or is the process final? Should I keep waiting and remain hopeful? I sincerely hope for a fair outcome. Thank you.
@CVPR Thank you for all your hard work on #CVPR2025. I’ve seen the accepted papers and final oral/highlight decisions. I submitted an appeal because my AC rejected my paper despite high review scores, with a funny meta-review that seemingly contradicts CVPR’s review guidelines.
@CVPR@MonaJalal_ What is the recommended way to submit an appeal this year? I have already directly emailed all PCs, but I have not received a response yet. Thank you!🥺