3D Avatar Creation from Photos Prototype

**Freelance Computer Vision / 3D Developer Needed — Self-Hosted 3 Photo to Fully Rigged 3D Avatar Prototype**

I am looking for an experienced developer to build a **self-hosted prototype** that generates a **fully rigged 3D face/head avatar from 3 photos of a real person**.

---

### Project goal

A user uploads:

* 1 front face photo
* 1 left-side face photo
* 1 right-side face photo

The system should generate:

* a **3D face/head avatar**
* high similarity to real person
* **fully rigged face for animation**
* **lip-sync ready setup**
* **eye movement support**
* export as **GLB (preferred)** or **FBX**

---

### Animation requirements (important)

The avatar must include:

* **basic facial rigging or blendshapes**
* mouth movement support for speech (lip-sync ready)
* eye movement / blinking capability
* ability to later drive animations via phonemes or visemes

(Does NOT need advanced cinematic animation — just functional for real-time app use)

---

### Scope

I do **not** need:

* mobile app UI
* voice generation
* full backend production system
* multiplayer / game engine integration

A working prototype + code is enough.

---

### Must be self-hosted

I do **NOT** want third-party paid APIs or hosted avatar services.

I want:

* fully **open-source based pipeline**
* runs on **my own machine/server**
* reusable in my own app later

---

### Preferred technical skills

* PyTorch / PyTorch3D
* OpenCV / MediaPipe
* Blender (very important for rigging)
* DECA / EMOCA / 3DDFA-V2 or similar face reconstruction models
* experience with facial rigging / blendshapes / animation pipelines

---

### Required deliverables

* full **source code**
* setup instructions (`requirements.txt` or Docker optional)
* clear README explaining pipeline
* runnable command like:

`3 photos → 3D avatar → rigged + animation-ready model`

Example:
`python generate_avatar.py --front front.jpg --left left.jpg --right right.jpg`

* exported **GLB or FBX**
* includes **rig + facial animation controls (eyes, mouth, basic expressions)**
* at least **2–3 sample outputs**

---

### Ownership

Upon final payment:

* full ownership of all custom code
* disclosure of all open-source models used
* disclosure of license restrictions for commercial use
* code must be reusable for integration into my own application

---

### Budget

Up to **$1,000 USD** for this project.

---

### When applying, please include:

* relevant portfolio / GitHub
* examples of:

* 3D face reconstruction OR
* rigging / blendshape work OR
* avatar animation systems
* estimated timeline
* confirmation you can handle **rigging + basic facial animation setup**