poster and slides on Google drive: https://drive.google.com/drive/folders/1FefDClzqGG3ae_49_S0vL0LamCKu64bF
Landmark Searching Based on Hand-Painted Building Lines
Google ML Winter Camp 2020 Project
We use three models to realize the process of image reconstruction, object detection and image similarity comparison.
Model: CycleGAN
Model: Retinanet
+ FCOS
Model: reconstruct imgs + original imgs -> autoencoder
# place models
cd webapp
python app.py # demo will run on http://127.0.0.1:5000/