Sign in

Nodes Browser

ComfyDeploy: How ComfyUI-Stereopsis works in ComfyUI?

What is ComfyUI-Stereopsis?

This initiative represents a solo venture dedicated to integrating a stereopsis effect within ComfyUI (Stable Diffusion). Presently, the project is focused on the refinement of node categorization within a unified framework, as it is in the early stages of development. However, it has achieved functionality in a fundamental capacity. By processing a video through the Side-by-Side (SBS) node and applying Frame Delay to one of the inputs, it facilitates the creation of a stereopsis effect. This effect is compatible with any Virtual Reality headset that supports SBS video playback, offering a practical application in immersive media experiences.

How to install it in ComfyDeploy?

Head over to the machine page

  1. Click on the "Create a new machine" button
  2. Select the Edit build steps
  3. Add a new step -> Custom Node
  4. Search for ComfyUI-Stereopsis and select it
  5. Close the build step dialig and then click on the "Save" button to rebuild the machine

ComfyUI-Stereopsis

This initiative represents a solo venture dedicated to integrating a stereopsis effect within ComfyUI (Stable Diffusion). By processing a series of sequential images through the Side-by-Side (SBS) node and applying Frame Delay to one of the inputs, it facilitates the creation of a stereopsis effect. This effect is compatible with any Virtual Reality headset that supports SBS video playback, offering a practical application in immersive media experiences.

Side By Side Module for Image Concatenation

Overview

<p align="center"> <img width="1723" alt="Side-by-Side" src="https://github.com/IsItDanOrAi/ComfyUI-Stereopsis/assets/162214102/c4b70e1e-56ff-4010-954c-da2371b87903"> </p>

The Side_by_side module is a utility designed for concatenating two images side by side into a single image. It is implemented in Python using the PyTorch library, making it suitable for both academic research and industrial applications where image manipulation and processing are required.

Features

Image Concatenation: Combines two images horizontally, aligning them side by side. Input Flexibility: Accepts two images as input, with no restriction on the images' width, but requires them to have the same height and channel depth. PyTorch Integration: Utilizes PyTorch for image manipulation, ensuring compatibility with other PyTorch-based pipelines and operations.

Technical Details

Input Types: The module requires two inputs, each labeled as left_image and right_image. Both inputs are expected to be images, represented as PyTorch tensors. Return Types: The output is a single image, which is the result of concatenating the input images side by side. Functionality: By leveraging the torch.cat function, it combines the input images along the width dimension, creating a seamless side-by-side image composition.

Requirements

PyTorch: As a PyTorch-based module, it requires an environment where PyTorch is installed and configured.

Frame Delay Module for Image Sequences

Overview

<p align="center"> <img width="677" alt="Frame Delay" src="https://github.com/IsItDanOrAi/ComfyUI-Stereopsis/assets/162214102/a8768fa4-596e-4484-af9c-9bbda21da0f5"> </p>

The FrameDelay module offers a sophisticated solution for manipulating image batches by repeating and inserting frames, thereby introducing a delay effect without altering the overall batch size. This functionality is crucial in applications requiring temporal adjustments within image sequences, such as video processing, animation, and dynamic visual effects creation. Features

Frame Selection and Delay: Allows the selection of a specific frame within a batch and introduces a delay by repeating the selected frame a specified number of times. Batch Size Maintenance: Ingeniously maintains the original batch size after frame insertion, ensuring consistency and compatibility with subsequent processing stages. Customizable Parameters: Provides flexibility in selecting the frame to delay and the extent of the delay, with sensible defaults and bounds.

Technical Details

Input Types: Accepts an image batch as input along with parameters to select the frame (selected_frame) and define the delay (frame_delay). Return Types: Outputs an image batch where the selected frame has been repeated according to the specified delay, with the batch size kept constant. PyTorch Integration: Built with PyTorch, this module seamlessly integrates into workflows that utilize PyTorch for image manipulation and tensor operations.

Requirements

PyTorch: As it relies on PyTorch for all tensor operations, having PyTorch installed and configured in the working environment is a prerequisite.