3D virtual staining with 2D and 2.5D U-Nets

This repository contains the ML and DL projects I implemented during my Bachelor's degree. It reflects the knowledge I gained from books and online sources. The models are trained on problems in NLP, Computer Vision, and more. This repository highlights my learning journey and the variety of projects I completed during my Bachelor's studies.

Voiceolation is a neural network for trained to extracts vocals from songs.

We segmented the Brain tumor using Brats dataset and as we know it is in 3D format we used the slicing method in which we slice the images in 2D form according to its 3 axis and then giving the model for training then combining waits to segment brain tumor. We used UNET model for our segmentation.

U-NET used for semantic segmentation of cultivated land in remote sensing image

"Attention UW-Net: A fully connected model for automatic segmentation and annotation of chest X-ray" by Debojyoti Pal, Pailla Balakrishna Reddy, and Sudipta Roy.

Using deep learning a U-net architecture is used to make segmentation, detection, and extraction of the lower left third molar. The result of the proposed U-net is compared with Attention U-net and U-net++.

Testing performances of different models of DL on a small training dataset of Solar Cells to find defects

Translate Brain MRI Images of one type to another

Damaged image restauration using a GAN model with a U-Net with skip connection autoencoder as a generator

U-Net from Scratch for Brain Tumor Segmentation

Breast cancer is one of the most common causes of death among women worldwide. Early detection helps reduce the number of premature deaths. In the study, I am working on creating a convolutional neural network capable of identifying tumor areas within medical images (which were taken with ultrasound).

Grayscale image colorization using a U-Net CNN (with VGG-19) and perceptual loss.

Visual analysis of the amount of advertising in the urban environment using computer vision (U-Net + PSP-Net) || Визуальный анализ заполненности городской среды элементами рекламы с помощью компьютерного зрения (U-Net + PSP-Net)

Nucleia image segmentation with U-net...

Satellite image segmentation using deep learning U-Net model, built at Hack36

The aim of this project is to detect copy move forgery in image forgery detection. Based on Deep learning techniques(CNN, NN)

Semantic Image Segmentation task solution for the Oxford-IIIT Pet dataset using U-Net neural network, Hyperband and Transfer Learning. AGH 2023/24

Pothole Segmentation using a custom U-Net Segmentation model built and trained from scratch using TensorFlow and Keras