• Stars
    star
    143
  • Rank 257,007 (Top 6 %)
  • Language
    C#
  • License
    Apache License 2.0
  • Created about 3 years ago
  • Updated 5 months ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

HolisticMotionCapture is an application and package that can capture the motion of a person with only a monocular color camera and move the VRM avatar's pose, face, and hands.

HolisticMotionCapture

demo

HolisticMotionCapture is an application that can capture the motion of a person with only a monocular color camera and move the VRM avatar's pose, face, and hands.

Pose, face and hands can be moved simultaneously or individually.

You can also output the rendered video to other applications.

Install HolisticMotionCapture application

HolisticMotionCapture can be downloaded from release page.

For Windows

Download installer(HolisticMotionCapture-1.1.0setup.exe file) from 1.1.0 release page and execute it.

For Mac OS

Download installer(HolisticMotionCapture_installer.pkg file) from 1.1.0 release page and execute it.

Usage HolisticMotionCapture application

Select your VRM file

  • You can change to your vrm avatar from pull down.
  • Images are loaded from and save to C:/Users/<user name>/AppData/LocalLow/IKEP/HolisticMotionCapture/VrmFiles directory in Windows.
  • Images are loaded from and save to /Users/<user name>/Library/Application Support/IKEP/HolisticMotionCapture/VrmFiles directory in Mac OS.
  • You can load new vrm avatar from the New VRM button.
    • You can continue to use the vrm file is loaded once, if application is restarting.

Image Device

  1. Select source input camera device from Input Device pull down.
  2. (Option) Set the resolution of camera images in W and H input field.
  3. Push the Start/Stop button for starting or stopping camera capture.

BackGround Select

  • You can change the background image from pull down.
  • Images are loaded from and save to C:/Users/<user name>/AppData/LocalLow/IKEP/HolisticMotionCapture/LoadedImages directory in Windows.
  • Images are loaded from and save to /Users/<user name>/Library/Application Support/IKEP/HolisticMotionCapture/LoadedImages directory in Mac OS.
  • You can output black image if you select the None option from pull down.
  • You can load new images as the background image from the New image button. HolisticMotionCapture is supported PNG(.png) or JPG(.jpg, .jpeg) images.
    • You can continue to use the image is loaded once, if application is restarting.

Output Control

For Windows

Rendered images can be output as the virtual camera image if Output on/off toggle is on. You can show composited image in another applications when you select a camera named HolisticMotionCapture in another applications.

For Mac OS

Rendered images can be output as the Syphon image if Output on/off toggle is on. You can receive composited image in another applications compatible with Syphon.

Avatar Control

  • Mode: You can choose to move your avatar's pose, face, or hands.
  • Reset pose button: You can reset your avatar pose.
  • Move upper body only toggle: You can choose to move only the upper body or move the whole body. It is useful when only your upper body is shown in the camera such as video conferencing.
  • Look Camera toggle: You can choose whether your avatar should always look at the camera or not.
  • Mouse Operation
    • The operation UI can be displayed/hidden with a mouse click.
    • You can move the virtual camera by mouse dragging and mouse scrolling.

For Developers

HolisticMotionCapture provides a package to move VRM avatars independently of my application.

Install HolisticMotionCapture package

HolisticMotionCapture package can be installed by adding following sections to your manifest file (Packages/manifest.json).

To the scopedRegistries section:

{
    "name": "Keijiro",
    "url": "https://registry.npmjs.com",
    "scopes": [ "jp.keijiro" ]
},
{
  "name": "creativeikep",
  "url": "https://registry.npmjs.com",
  "scopes": [ "jp.ikep" ]
}

To the dependencies section:

"jp.ikep.holistic-motion-capture": "1.1.0"

Finally, the manifest file looks like below:

{
    "scopedRegistries": [
        {
            "name": "Keijiro",
            "url": "https://registry.npmjs.com",
            "scopes": [ "jp.keijiro" ]
        },
        {
            "name": "creativeikep",
            "url": "https://registry.npmjs.com",
            "scopes": [ "jp.ikep" ]
        }
    ],
    "dependencies": {
        "jp.ikep.holistic-motion-capture": "1.1.0",
        ...
    }
}

Usage demo HolisticMotionCapture package

using UnityEngine;
using HolisticMotionCapture;

public class Visualizer : MonoBehaviour
{
    // Animator of VRM avatar
    [SerializeField] Animator avatar;

    HolisticMotionCapturePipeline motionCapture;
    WebCamTexture webCam;

    void Start()
    {
        // Initialize
        motionCapture = new HolisticMotionCapturePipeline(avatar);

        webCam = new WebCamTexture("Your webcam name", width, height);
        webCam.Play();
    }

    void Update()
    {
        // Ability to operate avatars with textures only.
        // You can also specify some optional arguments.
        motionCapture.AvatarPoseRender(webCam);
    }
}

Demo image

Video for demo was downloaded from here

Dependencies

HolisticMotionCapture package uses the following packages:

  • HolisticBarracuda
  • UniVRM
    • HolisticMotionCapture includes source codes of UniVRM. It's same that you write below packages to your manifest.json.
      "com.vrmc.vrmshaders": "https://github.com/vrm-c/UniVRM.git?path=/Assets/VRMShaders#v0.108.0",
      "com.vrmc.gltf": "https://github.com/vrm-c/UniVRM.git?path=/Assets/UniGLTF#v0.108.0",
      "com.vrmc.univrm": "https://github.com/vrm-c/UniVRM.git?path=/Assets/VRM#v0.108.0"
      
    • HolisticMotionCapture is not yet compatible with VRM 1.0.

HolisticMotionCapture application also uses the following packages:

Author

IKEP

LICENSE

Copyright (c) 2023 IKEP

Apache-2.0

More Repositories

1

BlazePoseBarracuda

BlazePoseBarracuda is a human 2D/3D pose estimation neural network that runs the Mediapipe Pose (BlazePose) pipeline on the Unity Barracuda with GPU.
C#
199
star
2

HolisticBarracuda

HolisticBarracuda is the Unity Package that simultaneously estimates 33 pose, 21 per-hand, and 468 facial landmarks on the Unity Barracuda with GPU.
C#
105
star
3

SelfieSegmentationBarracuda

SelfieSegmentationBarracuda is a human segmentation neural network that runs the Mediapipe Selfie Segmentation on the Unity Barracuda.
C#
47
star
4

UniCamEx

UniCamEx is a virtual camera for MacOS that can display textures output from Unity.
Swift
38
star
5

VirtualBackgroundCamera

VirtualBackgroundCamera is a virtual camera application that allows you to use a virtual background in any applications.
C#
38
star
6

PoseDetectionBarracuda

PoseDetectionBarracuda is a human pose detecter that runs the Mediapipe Pose Detection neural network model on the Unity Barracuda.
C#
19
star
7

PoseLandmarkBarracuda

PoseLandmarkBarracuda is a human pose landmark detecter that runs the Mediapipe Pose Landmark neural network model on the Unity Barracuda.
C#
16
star
8

FaceTrackingSystem

FaceTrackingデータをUDP通信により、Unity(PC)に転送しPC側でもレンダリングするデモ
C#
3
star
9

GPUBasedVRPaint

GPGPUとシェーダーによるVRペイント
C#
3
star
10

PhysicsButterfly

C#
2
star
11

HnadTrackingWithOpneCV

OpenCVでのハンドトラッキングを実装したゲーム
Processing
1
star
12

camera-calibration

Python
1
star
13

DeepLab-Spout

Virtual background sysnthetic like zoom application using DeepLab & Spout
C++
1
star
14

ButterflyEffect

ShaderLab(Unityシェーダーの学習用リポジトリ)
C#
1
star
15

I_AM_MAN

ISCA2018作品
C#
1
star