EPEdit: Redefining Image Editing with Generative AI and User-Centric Design

EPEdit is a web-based image editing platform designed to bridge the gap between powerful generative AI models and accessible user interfaces. While traditional professional software like Photoshop offers deep control, it presents a steep learning curve; conversely, many existing AI-driven tools are either too simplistic or require resource-intensive fine-tuning. EPEdit aims to provide a middle ground by integrating a robust backend with a user-friendly front-end, enabling users to perform advanced edits through simple text prompts and mask-based interactions.

The system architecture follows a Model-View-Controller (MVC) pattern. The front-end, built with ReactJS, provides a canvas for users to draw, mask, and interact with images. The backend, powered by Express, manages API requests and communicates with a Python-based model server. This server hosts pre-trained, zero-shot Stable Diffusion models, which are preloaded to ensure faster execution times. By utilizing zero-shot algorithms, EPEdit avoids the need for costly and time-consuming model retraining, allowing for efficient, on-the-fly image manipulation.

User evaluations involving 24 participants demonstrate that EPEdit is highly effective for tasks such as object removal, background modification, and thematic collection design. Participants reported that the interface is intuitive and significantly easier to use than traditional professional software. Quantitative results indicate that EPEdit outperforms Photoshop in terms of ease of use and speed for the specific creative tasks tested, making sophisticated generative AI capabilities accessible to hobbyists and students without requiring extensive technical training.

As generative AI continues to evolve, the challenge shifts from model capability to practical usability. EPEdit demonstrates that by abstracting complex AI operations into a streamlined, browser-based workflow, developers can democratize high-end image editing. This approach reduces the barrier to entry for creative professionals and casual users alike, proving that advanced generative features can be delivered efficiently without sacrificing user experience or requiring expensive hardware.