Prashant-mahajan / Few-shot-Image-Generation

General Adversial Networks using Few shot learning

Few Shot Image Generation

Generating images using GAN: Using four input images

Methods used:

1. FIGR: Meta-training DCGAN with Reptile

2. Transfer learning and data augmentation with following GAN architectures:

• DCGAN
• cDCGAN
• InfoGAN

---

Dataset

Evaluation Metrics:

1. MSE (Mean Squared Error): MSE calculates the mean squared error between each pixels for the two images we are comparing.

---

2. SSIM (Structural SIMilarity): The SSIM index is a method for measuring the similarity between two images. The SSIM index can be viewed as a quality measure of one of the images being compared, provided the other image is regarded as of perfect quality. It looks for similarities within pixels; i.e. if the pixels in the two images line up and or have similar pixel density values.

SSIM puts everything in a scale of -1 to 1. A score of 1 means they are very similar and a score of -1 means they are very different.

---

3. BRISQUE (Blind/Referenceless Image Spatial Quality Evaluator): It extracts the point wise statistics of local normalized luminance signals and measures image naturalness (or lack there of) based on measured deviations from a natural image model. We also model the distribution of pairwise statistics of adjacent normalized luminance signals which provides distortion orientation information.