VisionMorph - Transforming Text to Visual Sketches

Facial sketch generation plays an important role in areas such as forensic investigations, digital art, character design, and human-computer interaction.However, traditional sketching methods rely heavily on trained artists, making the process time consuming, subjective, and difficult to scale.In this paper, we present VisionMorph, an AI-based facial sketch generation system that creates realistic facial sketches from structured textual descriptions.The system allows users to select and describe facial features such as face shape, eyes, nose, lips, hair, and expressions through a user-friendly interface.These selected attributes are converted into meaningful natural language prompts, which are then processed using a Stable Diffusion model fine-tuned with LoRA to generate sketch-style images.This supports both initial sketch generation and further refinement using image-to-image editing, enabling users to make gradual improvements to the generated sketches.Experimental results show that the proposed system produces consistent and high-quality facial sketches suitable for both forensic and creative applications, while significantly reducing manual effort and overall generation time.