Paper Notes of CVPR-0155

Depth-Based 3D Hand Pose Estimation: From Current Achievements to Future Goals

Abstract

This paper is a study (or a survey) concerning 3D hand pose estimation. I catch a light on its figures about some curves and tables about some data.

####Dataset

Table 1 is the data set used in their experiments.

Method Comparison

Table 2 is method comparison. Note that it included method name, model, input size, parameters range, some specific configurations.

Errors-Parms Figure

Figure 2 show the correlation between estimation errors and error threshold. Note that different method is represented with different curve or color.

Multi Analyze

For an important metric, we should analyze it in many ways. For example, its’ disciplinarian of changes under different parameters.

Local Analyze

Sometimes, we should not only analyze the whole, but also focus the individual in the whole.

Conclusion

This paper is a study on 3D hand pose estimation which is far from my field. Thus, I mainly summarize it from the perspective of writing.