ViCoW: A Dataset for Colorization and Restoration of Vietnam War Imagery

ViCoW is a dataset of 1,896 high-resolution image pairs extracted from four historically significant Vietnamese films set during the Vietnam War era. Each pair includes a color frame and its grayscale version (converted using the ITU-R BT.601 formula), supporting research in AI-based image restoration and colorization.

Many photos and film scenes from this period exist only in black and white, making them feel distant to modern audiences. ViCoW helps bridge this gap by enabling the development of models that bring historical visuals to life in color. Frames were extracted every 3 seconds from well-preserved footage and manually curated for diversity and cultural relevance.

ViCoW contributes to both technical research and cultural preservation, offering a valuable resource for training and evaluating deep learning models in digital heritage, education, and historical storytelling.