Add a New Object
================

.. role:: guimenu

.. role:: guiaction

.. role:: guioption


This tutorial guides you through the process of adding a new object primarily
using :doc:`AI Hub </ai_hub>`. The steps include registering a 3D mesh model,
preparing an object detection model, and configuring pose estimation and
grasping capabilities. Follow these steps to enable your robot to recognize,
locate, and manipulate new objects.

.. contents::
    :local:
    :depth: 2
    :backlinks: none
    :class: toc


Step 1. Register a 3D Mesh Model
---------------------------------


1.1 Via externally sourced mesh model
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
You can upload a pre-existing 3D mesh model of an object to the AI Hub using
:ref:`AI Hub: Data Management <ai_hub:data management>`. See :doc:`Tutorial: Upload A Mesh Model </tutorials/upload_mesh>`

.. seealso::
    For a programmatic approach to perform data management, refer to
    :ref:`API: Data Management Client <api:Data Management Client>`.

1.2 Via 3D scanning
^^^^^^^^^^^^^^^^^^^

Use :ref:`AI Hub: 3D Object Scanning (Scan Box only) <ai_hub:3D Object Scanning (Scan Box only)>`
to create a texturized 3D mesh model of your desired object(s).

.. seealso::
    For a programmatic approach to perform workspace/environment scanning,
    refer to :ref:`API: Robot Scan Client <api:Robot Scan Client>`.


Step 2. Prepare an Object Detection model
------------------------------------------

2.1 Generate a Training Dataset
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Use :ref:`AI Hub: Data Generation <ai_hub:Data Generation>` to generate one/several
dataset(s) consisting your desired object(s).

For detailed instructions on creating synthetic and/or real datasets, refer to:

    * :doc:`Tutorial: Generate Synthetic Dataset </tutorials/generate_synthetic_data>`
    * :doc:`Tutorial: Collect Real Data </tutorials/collect_real_data>`
    * :doc:`Tutorial: Annotate Real Data </tutorials/annotate_data>`

.. seealso::
    To programmatically generate a dataset, refer to the following client APIs:

    - :ref:`API: Data Generation Client <api:Data Generation Client>`
    - :ref:`API: Real Data Collection Client <api:Real Data Collection Client>`

2.2 Train an Object Detection Model
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Utilize :ref:`AI Hub: Model Training and Testing <ai_hub:Model Training and Testing>`
to train an object detection model on your registered dataset(s). See :doc:`Tutorial: Train an Object Detection Model </tutorials/train_object_detection>`

.. seealso::
    For a programmatic approach to train and infer an object detection model,
    go to :ref:`API: Instance Segmentation Client <api:Instance Segmentation Client>`

Step 3. Prepare a Pose Estimation model
---------------------------------------

The default pose estimation model ``general/v_1`` that uses the method
``neura_nonDLPE1`` is a deterministic pose estimation algorithm. After
registering the object mesh model, no additional action is required to use
this model to perform 6D object pose estimation.

.. attention::
    Currently, no pose estimation method that require training are available.

.. seealso::
    For a programmatic approach to infer a pose estimation model, go to
    :ref:`API: Pose Estimation Client <api:Pose Estimation Client>`

Step 4. Enable Grasping
-----------------------

To perform pose-aware picks in an application, one is required to record grasps
of the objects that will be picked. See :ref:`Tutorial: Enable Grasping of New Object <tutorials:Enable Grasping of New Object>` to learn how 
to teach grasp poses.