Home

Awesome

DSALVANet

This is the code for Few-Shot Object Counting with Dynamic Similarity-Aware in Latent Space.

<div align=center> <img src="./test_data/4297.jpg" width="300"/><img src="./output/output.jpg" width="300"/> </div> <div align=center> <img src="./test_data/P0787.png" width="250"/><img src="./output/output1.jpg" width="250"/><img src="./output/output2.jpg" width="250"/> </div>

Dependencies

We are good in the environment:

Installation

pip install -r requirements.txt

Pre-trained models

Usage

python test.py -w ./checkpoints/checkpoint_200.pth -i ./test_data/4297.jpg -b ./test_data/4297.txt

Others

We will release more details of DSALVANet after the paper is officially publiced in the journal.

Before that, if you have any questions, you can contact me via email: kadvinj@outlook.com.