This project is my start into performing human detection and tracking through computer vision techniques.
Human Detection
Pose Estimation
MediaPipe Pose Landmarker
I use a PostGre SQL database to store my detections and am working on linking detections together.
As a dancer, a primary application is to analyze dance videos. Some issues associated with dancing:
Changing angles of the body can make it harder to connect detections across time frames.
Occlusion (one dancer blocks another, an arm blocking the torso) can also confuse the identity of the detection.
This is a library of GNU Radio tools. Currently this includes:
Throughput measurement block
File Recorder block
Table block (uses JDSP with a Java Swing Frame and Table model)
JDSP is a Java library of common DSP algorithms. It is basically my attempt to provide basic functions I enjoy from Python (numpy, scipy, matplotlib) to development in the Java environment.
Things implemented:
Filter design (windowing), convolution
I/O (CSV, PCM Audio file)
Swing UI for plotting line, spectrogram, tables
py-boring-stuff is a Python library that can be used to map out a Python library. This can be used for documentation of your own library or as a tool to analyze external libraries. The main use-case is to use 'inspect' to map a Python library and then generate a class diagram in PlantUML format.
This is a Python library I created when I first was looking into how the OFDM is generated. I eventually added some parameter estimation and a receiver. The receiver is currently only supporting 802.11ah HaLow because there is a lot of information needed to receive/decode an OFDM signal (preamble, pilot locations and values, coding scheme).