Traditional Turkish rugs are closely interwoven into the country’s cultural heritage, reflecting centuries of imagination, and storytelling. Characterized by complicated compositions, brilliant colors, and symbolic motifs, these rugs present not only decoration but carry Turkish history and values. Each design has a meaning in the culture for portraying themes such as love, protection, and nature; at the same time, it can reflects specific regional varieties typical for Anatolia, Van, Hakkari. But other than that, rugs play an important role in daily life, reflecting and tradition; they have formed very important objects of trade between East and West over many centuries. This work paying homage to the rich history with a collection of designs for rugs created by Turkish craftsmanship, blending cultural conservation and innovation, ensuring timeless designs relevant to today’s world. The mathematical algorithms behind the rugs make them unique and special. Nowadays, with the development of technology and the industrial revolution, this craftmanship has disappeared. I am doing this project to remind people of this lost practice. In this project I focused on creating Van rugs with AI.
Objective:
- Aim of this project is to show how ancient art can be combined with the latest technology. The guiding question in this project is: Can artificial intelligence depict the details of traditional Turkish rug designs?
- The goal is to create an output in which the past and future meet. The motifs, colors and patterns of rugs are very important in Turkish culture. I want to see how artificial intelligence will adapt to the rug features and interpret them.
- To automate the design process for Turkish rugs using a generative AI model.
- To preserve traditional artistry while opening new possibilities for innovation in rug design.
- To raise awareness to a culture that has now disappeared
- “CompVis/stable-diffusion-v1-4” ai model is used to train. It is a text-to-image ai model. The model creates images from textual descriptions. With the latent diffusion process it ensures computational efficiency without compromising image quality.
- For training ai, a a csv file created, which includes filename, design name and an example prompt.
1. Data Collection:
- A collection of images of traditional Turkish rugs will be created, sourced from Van Kilims (1998) book. The collection will encompass rugs from different regions and eras, providing a diverse and comprehensive dataset for AI learning.
2. Preprocessing:
- A process of refinement that includes cleaning, resizing and adding description to images. This process is important to make the images understandable by AI.
3. AI Training:
- With the collection of Turkish rugs, AI trained to create a tool that creates traditional Van rugs.
- Stable Diffusion AI model used, text-to-image generation model.
- Fine-tuned the model on the Turkish rug dataset to enhance its understanding of rug designs.
4. Testing and Evaluating Results:
- The rugs will be evaluated not only on technical perfection but also its spirit. The outputs should be meaningful with its pattern and color. Additionally, AI’s creation will be evaluated by hoe they reflect the meanings and spirits of Turkish rugs.