🔥 Real-time Super Resolution enhancement (4x) with content loss and relativistic adversarial optimization 🔥

[Remote Sensing] PyTorch implementation for "Remote Sensing Change Detection Based on Multidirectional Adaptive Feature Fusion and Perceptual Similarity"

PyTorch Implementation for Paper "Toward Multimodal Image-to-Image Translation"

Predicting gradient (of perceptual loss) without doing a backward pass

Deep learning-based image restoration pipeline with DnCNN, NAFNet, and legacy joint models. Includes PSNR/SSIM/LPIPS evaluation and visual comparisons.

This repository features an image sharpening pipeline using Knowledge Distillation. A high-capacity Restormer acts as the teacher model, while a lightweight Mini-UNet is trained as the student to mimic its performance.

Shiu, Jia-Yin's Master Thesis Project: A super-resolution framework for license plate recognition based on deep learning

Implement a Generative Adversarial Network which is able to frontalize faces from the pose invariant features learned in our proposed pose attention guided profile to frontal network for extreme profile faces

Implementation of a Vision-Mamba network, integrating State Space Models (SSM) with a patch-based encoder–decoder for image inpainting, colorization, and denoising. Trained with L1, SSIM, and VGG perceptual losses to preserve both structure and perceptual realism.

Synthetic CT generation from NAC PET for PET/MR reconstrunction and attenuation correction

Neural style transfer pipeline built with PyTorch and VGG-19, applying artistic styles to content images via perceptual loss optimisation. Implements Gram matrix style loss, configurable content/style weighting, and quantitative output evaluation using SSIM and PSNR metrics. Designed with a clean CLI for reproducible experiments on CPU or GPU.

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