Explore projects

A pure Python document scanner built from scratch using only NumPy. Detects corners, corrects perspective, and binarizes images. No external libraries. Developed for academic purposes (TIPE project).

I developed a complete Human Gesture Recognition system based on AIoT technologies. I used the MetaMotionR wearable device, equipped with an accelerometer and a gyroscope, to collect real-time motion data. I then processed the raw signals through filtering, normalization, and segmentation techniques to create a clean and balanced dataset.

On top of this dataset, I trained machine learning models and neural networks capable of recognizing and classifying different hand and finger gestures with high accuracy. The final system enables automatic gesture recognition and can be applied in domains such as healthcare, industry, gaming, and automotive, providing natural and touchless interaction with technology.

Technical details:

MongoDB for storing, managing, and organizing the collected sensor data.

Data Engineering & Preprocessing: filtering (low-pass), normalization/standardization, segmentation into time windows, and optional dimensionality reduction with PCA.

Exploratory Data Analysis: visualization of raw and processed signals (accelerometer, gyroscope, combined 6-axis data) and dataset balance checking.

Machine Learning Models: applied both Support Vector Machines (SVM) for statistical learning and Neural Networks (mainly CNNs using TensorFlow) for deep learning.

Model Evaluation: confusion matrices, classification reports, and learning curves (accuracy, loss) to assess performance.

Hyperparameter Tuning: grid search and fine-tuning to improve model robustness.

Python Ecosystem: implemented with NumPy, SciPy, pandas, scikit-learn, TensorFlow, matplotlib, seaborn, ensuring reproducibility through a requirements.txt.

人工智能学习路线图，整理近200个实战案例与项目，免费提供配套教材，零基础入门，就业实战！包括：Python，数学，机器学习，数据分析，深度学习，计算机视觉，自然语言处理，PyTorch tensorflow machine-learning,deep-learning data-analysis data-mining mathematics data-science artificial-intelligence python tensorflow tensorflow2 caffe keras pytorch algorithm numpy pandas matplotlib seaborn nlp cv等热门领域

人工智能学习路线图，整理近200个实战案例与项目，免费提供配套教材，零基础入门，就业实战！包括：Python，数学，机器学习，数据分析，深度学习，计算机视觉，自然语言处理，PyTorch tensorflow machine-learning,deep-learning data-analysis data-mining mathematics data-science artificial-intelligence python tensorflow tensorflow2 caffe keras pytorch algorithm numpy pandas matplotlib seaborn nlp cv等热门领域

NumPy Practice Labs | This repo collects 69 of programming labs exercises for NumPy Practice Labs. This course contains lots of labs for NumPy, each lab is a small NumPy project with detailed guidance and solutions. You can practice your NumPy skills by completing these labs, improve your coding ...

Quick Start with NumPy | This repo collects 15 of programming labs exercises for Quick Start with NumPy. This course will teach you the fundamentals of NumPy, a library that supports many mathematical operations.

llama3.np is pure NumPy implementation for Llama 3 model.

My solutions for Advent of Code 2024 in Python. I use Numpy a lot, sorry.

Toolbox for analysis of magnetic system. The computation engine support multiple backend, e.g. NumPy and OpenCL. API is exported to python. Visualization tools are included.

Deep Learning Library from scratch using NumPy and SciPy

Big Data Beer Reccomender Dask Project. Uses python, numpy and dask. (DATA301)

A single-layer neural network implementation for MNIST digit classification using NumPy for deep learning (COSC440).