Ziqi Ye

Kernkompetenz: *Reglumgstechnik, Messtechnik, TM bis Ⅲ, Mathe bis Ⅳ, ETIT bis Ⅱ, DSS, CV, DL jetzige Tätigkeit: *Nachhilfe bei den ETIT und Mathe für Bachelor Studenten *STM32 lernen Fokus des letzten Semester: *Projekte in CVⅠsowie Machine learning (Edge Detection, Optical Flow, Object Recognition, Stereo Vision, Regression, KNN, PCA, SVC, K-Mean, NN, CNN...) Projekte: *Anomaly Detection in Pneumatic Cylinder Production " A supervised learning model were trained with labeled data from real procssing to predict the quality test result of cylinder bottom. The training data (stored as .h5 files) comes from sensor readings in cnc milling machine in TU Darmstadt. Statistic features were generated from the time series data. LDA were performed to find the most relevant features to reduce the dimensions of the data. With a stacked model of LDA + KNN, PCA + SVC, NCA + SVC, the total accuracy can reach over 99%, plus being still stable and robust. "