Niki Amini-Naieni
@nikigoliAI
Followers
16
Following
6
Media
6
Statuses
20
🚀 New paper out! I'm excited to share our latest work: "Open-World Object Counting in Videos", now on arXiv: https://t.co/RSK998xhmy. In this paper, we introduce CountVid — the first model for counting any object in videos — and VideoCount, a new dataset for this novel task.
1
0
1
Work with @TengdaHan and Andrew Zisserman at VGG Oxford. Paper: https://t.co/nGzTsFvSXl, app: https://t.co/LsWxrbTo1j, code:
github.com
Includes the code for training and testing the CountGD model from the paper CountGD: Multi-Modal Open-World Counting. - niki-amini-naieni/CountGD
0
0
1
"CountGD could represent one of the biggest advances in colonial wildlife monitoring in conservation history and it has already helped us get through large datasets” - Dr Tom Hart (Penguin Expert, Senior Lecturer)
0
0
0
Penguin researchers at Oxford Brookes University and Penguin Watch have been using it to count penguins in aerial images to better understand factors influencing their population size.
1
0
0
Annolid has an open-source counting tool powered by CountGD. Check out the YouTube video that showcases it here:
1
0
0
“When I came across CountGD, I was genuinely impressed. I think it’s a game-changer—even compared to Google Gemini and open-source models like Quen. It’s working like a magic tool for us. CountGD is revolutionary.” --Associate Director, software and firmware at Dhruva Space
1
0
0
Dhruva Space has been using CountGD for earth observation, to count objects from satellite images.
1
0
0
Saiwa has integrated CountGD into its Training Free Object Counting with Prompts online counting tool ( https://t.co/b0LmTHvFPb) "Saiwa's Online Counting Tool, built on the advanced CountGD framework, leverages its cutting-edge multi-modal object counting capabilities."
saiwa.ai
Training-free Object Counting with Prompts is revolutionizing the way we approach tasks in sectors ranging from retail and manufacturing to environmental m
1
0
0
@LandingAI has integrated CountGD into its VisionAgent tool ( https://t.co/cmnPHIgea0) “Our internal benchmarks revealed that the Multi-Modal Counting Model powering VisionAgent, CountGD, consistently delivers results with less than a 5% margin of error.”
1
0
0
I was recently really excited to find out many people have been successfully using our CountGD model (NeurIPS'24) for products, open-source tools and science applications.
2
1
6
work with @ronnieclark__ , @JakabTomas, and Andrea Vedaldi at PIXL and VGG, Oxford.
0
0
0
We present the first approach to uncertainty calibration for NeRFs! ✨ Our meta-calibrator calibrates NeRFs without any ground truth data from the target scene. Come say hi at the Wednesday afternoon poster session at @eccvconf! Read the paper here: https://t.co/yXF8CzfpsN
3
1
2
Our recent paper, CountGD: Multi-Modal Open-World Counting, has been accepted to NeurIPS'24! 🔥 CountGD can count given only text or only example images or both for any object. Try the app here: https://t.co/LsWxrbTo1j Read the paper here: https://t.co/nGzTsFvSXl
4
4
14