3D Image Composer

An interactive application that segments a person from an image, inserts them into a stereoscopic background, and creates an anaglyph 3D image. The project allows dynamic depth adjustment to position the person closer or farther in the scene.

---

Features

• Person Segmentation: Automatically detects and isolates a person in an image using a pre-trained deep learning model.
• Stereoscopic Integration: Inserts the segmented person into left and right stereoscopic images.
• Anaglyph 3D Images: Combines left and right images into an anaglyph viewable with red-cyan glasses.
• Dynamic Depth Adjustment: Adjust the perceived depth of the inserted person with a slider.
• Interactive Gradio App: User-friendly interface for uploading images and generating results.

---

Installation

Prerequisites

• Python 3.9 or later
• pip (Python package manager)
• Git

Steps

1. Clone the repository:

   git clone <repository-url>
   cd project

2. Install dependencies:

   pip install -r requirements.txt

3. Ensure the following directories exist:

   mkdir assets output

4. Add input images to the assets/ directory:

   • A person image (e.g., person.jpg).
   • A side-by-side stereoscopic image (e.g., stereo.jpg).

---

Running the Gradio App

1. Launch the app:

   python src/main.py

2. Open the provided URL in a browser.

3. Upload your images:

   • Person Image: Upload the image to segment.
   • Stereoscopic Image: Upload the side-by-side stereoscopic image.

4. Adjust the depth level using the slider to generate a 3D anaglyph.

---

Running Locally

Segment a Person

1. Run the segmentation script:

   python src/segmentation/segment_person.py

2. This generates the segmented image with a transparent background in the output/ directory:

   • output/person_segmented.png

Insert the Person into a Stereoscopic Image

1. Run the insertion script:

   python src/stereoscopic/insert_person.py

2. This generates modified left and right stereoscopic images:

   • output/left_with_person.png
   • output/right_with_person.png

Generate Anaglyphs

1. Run the anaglyph creation script:

   python src/anaglyph/create_anaglyph.py

2. This creates anaglyph images for different depths:

   • output/anaglyph_close.png
   • output/anaglyph_medium.png
   • output/anaglyph_far.png

---

Outputs

Generated Files

• Segmented Person:
  • output/person_segmented.png (Transparent image of the detected person)
• Stereoscopic Pairs:
  • output/left_with_person.png
  • output/right_with_person.png
• Anaglyph Images:
  • output/anaglyph_close.png
  • output/anaglyph_medium.png
  • output/anaglyph_far.png

Example

The resulting anaglyph images can be viewed with red-cyan 3D glasses to see the depth effect.

---

Future Enhancements

• Support for multiple objects with independent depth adjustments.
• Enhanced background blending for better integration.
• Integration with other 3D visualization techniques (e.g., VR).
• Real-time updates with WebSocket support.

---