Term: Spring 2019

• Team Group 6

• Team members

  • Jingwen Wang ([email protected])
  • Weixuan Wu ([email protected])
  • Ying Jin ([email protected])
  • Yuqiao Li ([email protected])
  • Ziyi Liao ([email protected])

• Project summary:

In this project, we created a classification engine for enhance the resolution of images. We conducted the Feature Extraction and used GBM model to complete the Baseline Model. After tuning model parameter, the baseline model can reach accuracy at 99% level. In order to facilitate the computational efficiency, we develop an enhanced model by using XGboost method. This method greatly reduces the training and testing time and it reaches at 95% accuracy level. To further improve the image, we use superResolution technique to enlarge the image. As a result, images conducted by our model become smoother comparing to those are enlarged directly.

• Baseline_0 (with depth = 3, 5, 7, 9, 11 and n.trees=200)

• Baseline_improved(with depth = 9,11, shrinkage = 0.1, 0.01, and 0.001,n.trees=1000)

• Extreme Gradient Boosting (with depth=6,7,8 and eta=0.5,0.6,0.7,0.8 and 0.9)

Our team also explore other methods in order to improve the training process. We tested SVM and Random Forest on small samples. These two methods cost relatively long time even for small smaples. We are still looking for solutions on applying these two methods on large training data set.

• SVM with RBF kernel (with cost = 0.1,1,10,50 and gamma = 0.01,0.1,1,10)

• Random Forest (n_tree = 200, training error: 0.28%)

Contribution statement: (default) All team members contributed equally in all stages of this project. All team members approve our work presented in this GitHub repository including this contributions statement.

• Jingwen Wang(jw3667): Worked on Baseline and Baseline_improved model; wrote the superResolution function; collaborated with Ziyi Liao, completed the main.Rmd; helped with writing ReadMe files.

• Yuqiao Li(yl3965): Worked on SVM model with RBF kernel; wrote train_svm.Rmd and cross_validation_svm.Rmd; tested the model on a small sample set. Wrote training result of small samples in notes_svm.Rmd. Helped with ReadMe.

• Ying Jin(yj2453): Worked on feature extraction; wrote feature.R. Prepare the presentation and slides; helped with ReadMe.

• Ziyi Liao(zl2739): Worked on Extreme Gradient Boosting(XGBoost model); collaborated with Jingwen Wang, completed the main.Rmd; prepare the presentation and slides. Helped with ReadMe.

• Weixuan Wu(ww2493): Worked on feature generation; wrote train_rf.R, test_rf.R, and cross_validation_rf.R, and tested the model on a small sample (200 images).

Following suggestions by RICH FITZJOHN (@richfitz). This folder is organized as follows.

proj/
├── lib/
├── data/
├── doc/
├── figs/
└── output/

Please see each subfolder for a README file.