This Repository Contains Solution to the Assignments of the Generative Adversarial Networks (GANs) Specialization from deeplearning.ai on Coursera Taught by Sharon Zhou, Eda Zhou, Eric Zelikman
Redesigning the Pix2Pix model for small datasets with fewer parameters and different PatchGAN architecture
✨ Dive into image denoising magic! This project uses Attention U-Net and PatchGAN to tackle noise types like low Gaussian and salt-and-pepper noise. Perfect for computer vision, deep learning, and generative modeling enthusiasts. Restore clarity to noisy images with cutting-edge AI! 🚀🎨
Using Pix2Pix GAN for translating Anime images to something more aesthetic
Generate Faces Using Deep Convolutional Generative Adversarial Networks (DCGAN)
A PyTorch-based image synthesis and computer vision framework.
Sketch-to-Image Criminal Identification System using a pix2pix GAN trained on the CUHK dataset to generate realistic images from sketches. A classifier is then trained on mugshot data to identify individuals based on generated images, enhancing forensic sketch recognition with deep learning for accurate identification.
A PyTorch implementation of the Pix2Pix model (cGAN) for image-to-image translation. This project translates semantic segmentation maps to photorealistic images using the Cityscapes dataset.
An Image colorization algorithm using PatchGan and Convolution Block Attention Modules (CBAM)
Face Aging & De-Aging using CycleGAN, PyTorch, OpenCV and Gradio. AI-powered facial age transformation with an interactive web interface.
An end2end PyTorch CycleGAN framework for unpaired facial age translation, with full training pipeline, AWS/S3 utilities, and an interactive Gradio demo.
Analysis of different models for mobile ocular biometrics.
Colorizing grayscale facial images using a U-Net neural network architecture with Patch GAN
Generating Maps from Satellite images using the Pix2Pix GAN.
Implementing and comparing CGAN and WGAN-GP for grayscale image colorization using U-Net and PatchGAN in L*a*b color space, with experiments on ImageNet and MS-COCO.
A PyTorch implementation of Image-to-Image Translation with Conditional Adversarial Networks (Isola et al., CVPR 2017), applied to the anime sketch-to-colour task: given a black-and-white sketch of an anime character, the model learns to produce a plausible full-colour version.
Task: Neural Style Transfer. The implemented solution uses a CycleGan architecture.
To associate your repository with the patchgan topic, visit your repo's landing page and select "manage topics."