Method partially based on the excellent work described in:
[1] Song, Qingyu, et al. "Rethinking counting and localization in crowds: A purely point-based framework." Proceedings of the IEEE/CVF International Conference on Computer Vision. 2021. [2] Khirodkar, Rawal, Shashank Tripathi, and Kris Kitani. "Occluded human mesh recovery." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022. [3] Cheng, Jian, et al. "Decoupled two-stage crowd counting and beyond." IEEE Transactions on Image Processing 30 (2021): 2862-2875.
Terms of Service
All images are compared against multiple databases of illegal images such that any match against this database will be reported to the appropriate authorities.
No warranty expressed or implied is provided as to the functioning of this service.
Attempting to process sexually explicit, violent, hateful or other objectionable material may result in a site-wide ban from this service.
Privacy Policy
All images are stored temporarily server-side insofar as to complete the image or video processing task.
When processing is complete, both the input and processed image are irretrievably destroyed.
When processing is complete, the resulting image or video is stored for (1) one hour before being irretrievably destroyed.
⟶
