ColorNet: Generating RGB colors from text sequences using RNNs

A neural network that predicts RGB colors from text descriptions. Give it color names or descriptive phrases like "sunset orange" or "ocean blue", and it generates corresponding RGB values.

How it Works

• Uses LSTM layers to process text input character by character
• Converts text descriptions into numerical sequences
• Outputs normalized RGB values between 0-1
• Trained on a dataset of color names and their RGB values

Examples

Input -> Output (RGB)

• "ocean blue" -> (41, 128, 185)
• "forest" -> (34, 139, 34)
• "sunset orange" -> (253, 94, 83)
• "storm grey" -> (119, 136, 153)

Usage

1. Install dependencies:

pip install -r requirements.txt

2. Run the Streamlit app:

streamlit run app.py

Wanna try it out in your browser? Visit at https://colornet.streamlit.app/

Requirements

• Python 3.7+
• TensorFlow 2.x
• Pandas
• NumPy
• Matplotlib
• Streamlit