AI-BASED VIDEO UPSCALER

This project focuses on enhancing video quality by upscaling lower-resolution videos to higher resolutions using advanced AI techniques. We utilized convolutional neural networks (CNNs) and generative adversarial networks (GANs) to improve video clarity and detail, resulting in high-resolution video. The project showcases both static images and videos demonstrating the effectiveness this approach.

Techniques Used

Convolutional Neural Networks (CNNs): CNNs are employed to extract and enhance features from low-resolution frames, making them suitable for capturing fine details and textures.
Generative Adversarial Networks (GANs): GANs are used to generate high-resolution frames by training two neural networks—the generator and the discriminator—in a competitive manner, leading to sharper images.
Super-Resolution Convolutional Neural Network (SRCNN): This specialized type of CNN focuses on super-resolution tasks, reconstructing high-resolution images from low-resolution inputs.
Perceptual Loss: A technique that uses features from pre-trained deep networks to guide the training process, ensuring that the upscaled images not only look sharp but also retain the original details.