Codeczh/FGVC-IoT

Introduction

This is the source code for our paper Robust Fine-Grained Visual Recognition with Images Based on Internet of Things

Network Architecture

The architecture of our proposed DDN model is as follows

Installation

After creating a virtual environment of python 3.5, run pip install -r requirements.txt to install all dependencies.

How to use

The code is currently tested only on GPU.

• Data Preparation

  Download data into DDN root directory and uncompress them using

  wget https://fgdata-sh.oss-cn-shanghai.aliyuncs.com/CUB200-Google-100-DDN.tar.gz
  wget https://fgdata-sh.oss-cn-shanghai.aliyuncs.com/CUB200-Google-100-Bilinear.tar.gz
  tar -xvf CUB200-Google-100-DDN.tar.gz
  tar -xvf CUB200-Google-100-Bilinear.tar.gz
  

• Demo

  If you just want to do a quick test on the model and check the final fine-grained recognition performance, please follow subsequent steps

  • Download one of the following trained models into demo/model using

    wget https://fgdata-sh.oss-cn-shanghai.aliyuncs.com/Model/DDN_Performance_79.92.pth
    

    Model	Description	Performance(%)
    DDN_Performance_79.92.pth	leverages 100 web images for each category (DDN)	79.92

  • Rename the downloaded model to step2_vgg_16_epoch_best.pth

  • Create Test data symbolic in demo by ln -s ../CUB200-Google-100-Bilinear data

  • Activate virtual environment(e.g. conda)

  • Run demo code using CUDA_VISIBLE_DEVICES=0 python demo.py

• Source Code

  If you want to train the whole network from beginning using source code, please follow subsequent steps (please make sure you have at least 200GB space)

  • Modify CUDA_VISIBLE_DEVICES to proper cuda device id in runall.sh,
  • Activate virtual environment(e.g. conda) and then run the script

    bash runall.sh