Update Readme with dataset instructions.

Readme.md

@@ -40,8 +40,33 @@ $ python inference.py --cfg configs/inference/inference.yaml
                       --skip_optimizer
 ```
 
-3\. To inspect the gaze tracking results run the jupyter notebook in `notebooks/view-inference_results.ipynb`
+3\. To inspect the gaze tracking results run the jupyter notebook in `notebooks/view-inference_results.ipynb`.
 
 Bash scripts for the above commands can be found in the `scripts` directory.
 
+## Prepare Datasets 
+
+### ETH-XGaze
+1\. Download the 448*448 pixels version of the datset from the official source [here](https://ait.ethz.ch/xgaze). Place the dataset in the `datasets` folder in the root of this directory.
+
+2\. Fit 3D eyes on iamges using the following command. This will export a `.pkl` data file in the dataset's folder which is used for training.
+```
+cd tools
+python xgaze_preprocess.py
+```
+
+3\. To visualize the 3D eye fittings run the notebook in `notebooks/xgaze_view_dataset.ipynb`.
+
+### Gaze360
+
+1\. Download the datset from the official source [here](http://gaze360.csail.mit.edu/). Place the dataset in the `datasets` folder in the root of this directory.
+
+### GazeCapture
+
+1\. Download the datset from the official source [here](https://gazecapture.csail.mit.edu/). Place the dataset in the `datasets` folder in the root of this directory.
+
+### MPIIFaceGaze
+
+1\. Download the datset from the official source [here](https://www.perceptualui.org/research/datasets/MPIIFaceGaze/). Place the dataset in the `datasets` folder in the root of this directory.
+