Deep Learning2020 – 2021ML Engineer

Neural Style Transfer

Inception v4 artistic transfer between image pairs

Neural-style transfer using Inception v4 so two images can be combined — framed as a tool for artists mixing historical references with recent work.

The Problem & Engineering Constraint

The Core Challenge

Style transfer needs a deep feature backbone that can separate content from texture without a custom dataset per artist.
Technical Architecture & Approach

Engineering Solution & Implementation

Inception v4 in TensorFlow with SciPy / Matplotlib / NumPy around the transfer loop.

View repository on GitHub

Measured Production Impact

Verified Outcomes & Deliverables

Transfer pipeline for pairing a content image with a style image.

Technologies & Components

System Tooling & Technologies

PythonTensorFlowInception v4SciPyMatplotlibNumPy