Color Recognition for Rubik's Cube Robot

Shenglan Liu,Dong Jiang,Lin Feng,Feilong Wang,Zhanbo Feng,Xiang Liu,Shuai Guo,Bingjun Li,Yuchen Cong

DOI: https://doi.org/10.48550/arXiv.1901.03470

2019-01-11

Abstract:In this paper, we proposed three methods to solve color recognition of Rubik's cube, which includes one offline method and two online methods. Scatter balance \& extreme learning machine (SB-ELM), a offline method, is proposed to illustrate the efficiency of training based method. We also point out the conception of color drifting which indicates offline methods are always ineffectiveness and can not work well in continuous change circumstance. By contrast, dynamic weight label propagation is proposed for labeling blocks color by known center blocks color of Rubik's cube. Furthermore, weak label hierarchic propagation, another online method, is also proposed for unknown all color information but only utilizes weak label of center block in color recognition. We finally design a Rubik's cube robot and construct a dataset to illustrate the efficiency and effectiveness of our online methods and to indicate the ineffectiveness of offline method by color drifting in our dataset.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

This paper aims to solve the problems encountered by Rubik's cube robots in recognizing the colors of Rubik's cubes, especially the color drifting problem. Color drifting refers to inaccurate color recognition caused by environmental factors (such as changes in illumination, differences in materials, etc.), which makes offline methods ineffective in continuously changing environments. To this end, the author proposes three methods to solve the color recognition problem of Rubik's cubes:

1. **Offline method**:
- **Scatter Balance & Extreme Learning Machine (SB - ELM)**: This is a method based on supervised dimensionality reduction and classification. By introducing angle measurement, the within - class scatter matrix and the between - class scatter matrix are re - defined to improve the robustness of the model.

2. **Online method**:
- **Dynamic Weight Label Propagation (DWLP)**: Using the color information of the known center blocks, the labels are propagated by dynamically adjusting the weights, thereby recognizing the colors of other blocks.
- **Weak Label Hierarchic Propagation (WLHP)**: Based on the weak - supervision assumption, using the color information of the center blocks as a starting point, the colors of other blocks are recognized through hierarchical propagation.

To verify the effectiveness of these methods, the author constructed a data set containing 348 images and designed a Rubik's cube robot for experiments. The experimental results show that the offline method performs well in a single environment but has a poor effect in a changeable environment; while the online methods, especially DWLP, show a higher recognition rate in various environments.