YoloLSTM: Image-based Localization model

"Demo: Image-based Indoor Localization using Object Detection and LSTM" (ACM MobiSys 2024 Demo)

Setup

• Python 3.11

• Torch 2.0.1+, Torchvision 0.15.2+ and CUDA 11.7+

pip install torch==2.0.1+cu117 torchvision==0.15.2+cu117 --extra-index-url https://download.pytorch.org/whl/cu117

• Installing Packages

pip install -r requirements.txt

• Change branch to i208_regression. The master branch cannot be used.

git checkout i208_regression

• Place the i208 laboratory dataset in the root data_all/train, data_all/valid and data_all/test.

data_all/
├── train/
├── valid/
└── test/

• Start learning with the following command.

python main.py

Evaluation

You can compare with other models by setting models/<model_name>/state in config.py to True/False.

• PoseNet: Implementation of PoseNet: A Convolutional Network for Real-Time 6-DOF Camera Relocalization [Kendall2015ICCV]
• PoseLSTM: Implementation of Image-Based Localization Using LSTMs for Structured Feature Correlation [Walch2017ICCV]
• SimpleCNN: CNN with the same number of layers as YoloLSTM using the whole image as input