This is an implementation of Recurrent BIM-PoseNet for camera pose regression related to our upcoming work. More details will be available soon.
The training and the test data can be found in this repository.
Compatibility: Tensorflow 1.9.0, Cudatoolkit=9.0, Keras 1.2.2, Python 2.7
The initial weight file (GoogleNet V1 trained on the Places dataset) and the fine-tuned model weights can be found here. The following are the details of the fine-tuned weight files:
- SynCar - Weights of model fine-tuned on Synthetic Cartoonish images.
- SynPhoReal - Weights of model fine-tuned on Synthetic photo-realistic images.
- SynPhoRealTex - Weights of model fine-tuned on Synthetic photo-realistic textured images.
- GradmagSynCar - Weights of model fine-tuned on synthetic gradmag of SynCar images.
- EdgeRender - Weights of model fine-tuned on Synthetic edge render images.
Other details in the name of the weight files describes the parameters, such as window length, learning rate, batch, ...., etc.
Related publications:
- Acharya, D., Singha Roy, S., Khoshelham, K. and Winter, S., 2020. A Recurrent Deep Network for Estimating the Pose of Real Indoor Images from Synthetic Image Sequences. Sensors, 20(19), p.5492.
- Acharya, D., 2020. Visual indoor localisation using a 3D building model (Doctoral thesis, The University of Melbourne, Melbourne, Australia). Retrieved from http://hdl.handle.net/11343/239180
- Acharya, D., Khoshelham, K., and Winter, S., 2019. BIM-PoseNet: Indoor camera localisation using a 3D indoor model and deep learning from synthetic images. ISPRS Journal of Photogrammetry and Remote Sensing. 150: 245-258.
- Acharya, D., Singha Roy, S., Khoshelham, K. and Winter, S. 2019. Modelling uncertainty of single image indoor localisation using a 3D model and deep learning. In ISPRS Annals of Photogrammetry, Remote Sensing & Spatial Information Sciences, IV-2/W5, pages 247-254.