Related Experiment Video
Updated: Oct 19, 2025
06:48
Author Spotlight: Advancements in 3D Optical Imaging for Comprehensive Body Composition Assessment in Modern Research
Published on: June 7, 2024
1.6K
Study on 3D Clothing Color Application Based on Deep Learning-Enabled Macro-Micro Adversarial Network and Human Body
Jingmiao Liu1, Yu Ren1,2, Xiaotong Qin3
1General Graduate School of Keimyung University South Korea, Daegu 42601, Republic of Korea.
Computational Intelligence and Neuroscience
|September 20, 2021
Summary
This study introduces a deep learning model for realistic 3D clothing color display in virtual fitting systems. The new method accurately renders clothing colors, improving virtual try-on experiences.
Area of Science:
- Computer Vision
- Virtual Reality
- Digital Fashion
Background:
- Online shopping convenience drives demand for virtual fitting systems.
- Current virtual fitting systems struggle with accurate clothing color representation.
- This limitation hinders realistic virtual try-on experiences.
Purpose of the Study:
- To develop a 3D clothing color display model using deep learning.
- To enhance the accuracy and realism of virtual fitting systems.
- To support human modeling-driven applications.
Main Methods:
- Utilized a deep learning-based macro-micro adversarial network (MMAN) for image analysis.
- Implemented preprocessing steps on analyzed image data.
- Constructed a 3D model with original image color fidelity using UV mapping.
Main Results:
- Achieved high accuracy (0.972) with the MMAN algorithm.
- Generated 3D models with clear and accurate clothing color representation.
- Color difference within 0.01 compared to original images.
- Subjective volunteer evaluations exceeded 90 points.
Conclusions:
- Deep learning effectively creates 3D models with accurate original clothing colors.
- The proposed model significantly improves virtual fitting system realism.
- This research offers valuable insights for character modeling and simulation.
