- With this Kaggle Dataset NYC Taxi Trip Duration Competition, I use some scipy model to analyze to get RMSLE (Root Mean Squared Logarithmatic Error) Score. (0.4678)
Documents
- Pandas
- Numpy
- Seaborn
- Matplotlib
- Xgboost
- Scikit-learn (sklearn)
-
You can also look this full report in this link (/NYC Taxi Duration Analysis/Taxi Dataset Analysis.docx)
-
Result
2. Automatic Face Replacement Service for portrait rights protection using Face Recognition and cycleGAN
- With Face Recognition and CycleGAN model, my project group developed Automatic Face Replacement Service.
- This service detect human faces in given input picture, and then we used face recognition and recognize the known face which is featured with the default picture using FaceNet.
- This service give users back a revised picture that unknown faces in the picture are replaced with Simpson-like faces using cycleGAN.
- Using Google Cloud Platform, study CycleGAN model with 2 image (Human face, Simpson face) and get results (see below).
- And also, make web-page for demo in Django, Bootstrap 4.
Documents