This paper works towards finding an effective solution for the task of image feature manipulation using natural language commands. The authors aim to modify the relevant features of an image using a natural language description of the target image, such that the irrelevant features are not modified. Majority of the research in this domain focuses on generating completely new images using natural language description, and the few methodologies which attempt manipulation of existing images super in a number of aspects such as modification of irrelevant features or the aesthetic quality of the generated image. The authors propose an architecture that combines the best components of existing techniques to create an effective system to solve the stated task. The proposed architecture generates images at a high resolution to maintain the aesthetic quality of the image and ensures that the irrelevant content of the original image is not affected. The authors present a qualitative and quantitative analysis of the system as compared to the existing baselines and demonstrate the system for a relevant application such as virtual trial of clothes.