IEEE RAS SLAM Winter School Project
This repo is the code and documents of project 9d during IEEE RAS Winter School on SLAM in Deformable Environments. It has been done by several prospective students from both master program of Robotics at TU Delft and Autonomous Systems from DTU.
Brief Description
This work is to implement a Monocular Depth Estimation network using Tensorflow 2 and it is based on the previous work High Quality Monocular Depth Estimation via Transfer Learning with modifications.
Award
We have been awarded the 3rd Prize (6 groups over 19 groups were awarded 1st, 2nd, 3rd prize) and 100 AUD 
on IEEE RAS Winter School on SLAM in Deformable Environments. This repo contains the code and some documents about our project. If you have any problems, please feel free to contact weihaoxuan97 or siyuanwu99 at gmail dot com.
Tasks
- Finish the network. You need to configure environment, revise the dataset folder, adjust parameters, and fill this gap. The used functions may include: “tensorflow.keras.applications.DenseNet169”, “tf.keras.layers.Conv2D”, “tf.keras.layers.UpSampling2D”, “tf.keras.layers.Concatenate”, “tf.keras.layers.LeakyReLU”, and “tf.keras.Model”.
- Improve the loss function.
- Add some data augmentation.