Khawar Islam

Title: Khawar Islam and His Patent in AI-Driven Data Augmentation.

Introduction

Khawar Islam is an inventor based in Abu Dhabi, United Arab Emirates. His work includes a patented approach to creating enriched training images with diffusion models, hybrid image concatenation, and fractal blending.

Latest Patents

Islam is associated with 1 patent: “System and Method for Label-Preserving Data Augmentation with Diffusion Models.” The invention takes an original image and one or more prompts as inputs to a diffusion model, which generates a corresponding image. It then combines part of the original image with part of the generated image using a binary pixel-wise mask. The mask contains zero- or one-valued pixels and has at least one entire row or column set to zero. The method applies the mask and its complement to the respective images, then blends a fractal image into the resulting hybrid image to produce an augmented image.

Career Highlights

Islam is associated with Mohamed Bin Zayed University of Artificial Intelligence. His patent describes a method for generating varied training images while preserving labels, using a combination of diffusion-based image generation, masked concatenation, and fractal blending.

Collaborations

The listed coworkers associated with Islam include Muhammad Zaigham Zaheer and Arif Mahmood.

Conclusion

Khawar Islam’s patent presents a structured approach to AI-based data augmentation that combines generated imagery, mask-controlled image composition, and fractal blending to create enriched training images.