OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Image & Video AI

Neural Style Tf

Neural Style Tf is a Python implementation of neural artistic style transfer using TensorFlow, combining the content of an input image with the visual style of one or more reference images. Its documented techniques also cover video frames, color preservation, semantic segmentation and adjustable style blending.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Neural Style Tf — OpenViST project profile in Image & Video AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when you want a command-line, TensorFlow-based workflow for stylizing local image or video inputs and adjusting how styles are combined.

Who it is for

  • Developers and technical artists who want to run neural style transfer on their own image files.
  • Users experimenting with applying one or multiple artistic references to images.
  • People exploring style transfer for video frames or color-preserving image treatments.

Documented uses

  • Generate an image that combines a photograph’s content with an artistic reference.
  • Compare different style weights or blend multiple style images.
  • Process a video with an artistic style.
  • Explore a style-transfer result while preserving original colors.
Before you choose

Requirements & limits

The repository is public and identifies its license as GPL-3.0.

  • Requires TensorFlow, OpenCV and a separately downloaded VGG-19 weights file.
  • The README presents CUDA and cuDNN as optional but recommended.
  • The basic image command lists PNG, JPG, PPM and PGM; the video command lists MP4, MOV and MKV.
  • Style-content weighting, multiple-style weighting and an original-colors option are documented.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Apply a style to a single image

Create a stylized image from one content image and one style image using the documented basic command.

Before you start

  • TensorFlow and OpenCV installed.
  • VGG-19 weights downloaded and the file imagenet-vgg-verydeep-19.mat copied to the project directory.
  • One content image and one style image.
  • The README lists .png, .jpg, .ppm and .pgm as supported image formats.

Steps

  1. Place the content image in ./image_input and the style image in ./styles.

  2. Run bash stylize_image.sh ./image_input/lion.jpg ./styles/kandinsky.jpg, replacing the example paths with your image paths.

Expected outcome

The command is documented for producing a stylized output image from the supplied content and style images.

Guide limitations

  • Image paths must not use ~ to represent the home directory; use relative or absolute paths.
  • CUDA and cuDNN are optional, recommended dependencies rather than required prerequisites.
Documentation 1 ↗
Source-based practical guide

Apply a style to a video

Submit a content video and style image to the repository’s documented video stylization command.

Before you start

  • TensorFlow and OpenCV installed.
  • VGG-19 weights downloaded and the file imagenet-vgg-verydeep-19.mat copied to the project directory.
  • One content video and one style image.
  • The README lists .mp4, .mov and .mkv as supported video formats.

Steps

  1. Place the content video in ./video_input and the style image in ./styles.

  2. Run bash stylize_video.sh ./video_input/video.mp4 ./styles/kandinsky.jpg, replacing the example paths with your file paths.

Expected outcome

The documented command submits the video and style image for video stylization.

Guide limitations

  • The README describes this workflow as video frames; it does not specify output packaging or playback details.
  • CUDA and cuDNN are optional, recommended dependencies rather than required prerequisites.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
cysmith/neural-style-tf · Official homepage title · retained in its source language
How we review sourcesSuggest a correction