/Multi-Task-Learning

Primary LanguagePython

Multi-Task-Learning

This research shows how to simultaneously train and execute the tasks of semantic segmentation and depth estimation using a multi-task deep learning algorithm (like HydraNet).

Architecture

image Source

Output

video

Instructions

  1. Download the NYU Depth V2 dataset from here and change the path of the dataset in the code.
  2. Install the necessary libraries from requirements.txt
  3. Run python train.py to train the model.
  4. To run inference on the trained model, run python inference.py

References

This project is based on the paper "Real-Time Joint Semantic Segmentation and Depth Estimation Using Asymmetric Annotations". Some of the code has been adapted from the official repository.