Image colorization and pencil sketch conversion using deep learning and Opencv

This paper presents a unified framework for image colorization and pencil sketch conversion, integrating deep learning techniques with classical image processing methods. The proposed system addresses two key challenges in computer vision: restoring realistic colors in grayscale images and generating hand-drawn style sketches from photographs. For colorization, a hybrid model combining Convolutional Neural Networks (CNNs) and Generative Adversarial Networks (GANs) is employed to predict chrominance channels in the CIELAB color space, producing visually consistent and natural results. For sketch generation, a lightweight OpenCV-based pipeline is designed, leveraging grayscale transformation, image inversion, Gaussian blurring, and dodge blending to deliver high-quality artistic renderings without the need for model training. The modular design allows both tasks to operate independently or sequentially, enabling flexible workflows for diverse applications such as photo restoration, digital art, and content creation. Experimental evaluation on a dataset of 6,000 paired images demonstrates the effectiveness of the proposed approach, achieving a balance between computational efficiency, structural accuracy, and visual appeal. This work highlights the potential of combining artificial intelligence with traditional computer vision to create scalable, user-friendly tools that bridge the gap between technical image enhancement and artistic creativity.