Name: Boguss
Type: User
Company: UTAS (University of Tasmania) Machine Learning engineer
Bio: "Legend Never Die"
Professional Programmer.
A.I. Programmer & Researcher.
Location: Hobart, TAS, Australia
Boguss's Projects
:metal: awesome-semantic-segmentation
image classification with patch-base inference & training algorithm is used to analyze quantitative coverage of beggiatoa.
crop labeled images according to xml files
Fascinating coverage calculation based on image classification.
visualise CNN filter of the last Conv2D layer during training process for each epoch
convert one channel black mask into coloured (RGB) mask (.png)
image segmentation using voting algorithm based on trained classification CNN weight. This methodology allows segmentation to consider context of neighbour pixels on its classification stage.
Cognitive test which is for diagnosing dementia.
Comprehensive solution for searching best suit model of regression and classification using a utility python package called 'lazypredict'.
csv to xml converter for Object detection Anotation (Tensorflow).
Convert tensorflow2 csv to YOLOv5 txt format.
diagnosis of early dementia using NLP machine learning and speech pathology methodology
Automatic butchery macro with YOLOv5
Deepfakes Software For All
Swap face between two photos.
This repository contains the source code for the paper First Order Motion Model for Image Animation
This project uses Python and OpenCV to calculate the lens distortion of a GoPro and use that information to remove the distortion.
python draw grid over image for seafloor traditional annotation purpose.
Simple GUI-Template for everyone! easy to use. kind descriptions. high scalability.
Implementation of Segnet, FCN, UNet , PSPNet and other models in Keras.
image preprocessing applied to Keras dataset generator which is trainable in Tensorflow Keras model and its visualisation to check input image.
Image processing GUI with various filtering option and adjustable parameters.
simple & basic image processing codes for data adjustment
Segmentation models by training publicly available bovine iris datasets, BovineAAEyes80, and comparing combinations of the state-of-art deep learning techniques.
java chat app (client and server) implementation without socket (but model with controller)
Simple json to csv converter
Legacy Management System for exclusive multi user server