This paper proposes a novel approach to the image colourization problem by fusing U-Net architecture with an Image Colourfulness Frame (ICF). The proposed method aims to overcome the limitations of existing colourization methods, which often produce pixelated and subpar photos. Colour maps are produced using the U-Net architecture using the grayscale input photographs. The colour maps' vibrancy is then evaluated using the ICF. The ICF helps to make colourization results more accurate and aesthetically pleasing. Experimental results show that the proposed method outperforms the state-of-the-art methods in terms of colourization accuracy and visual quality. The suggested method has many applications, including colourizing films and photos, restoring black and white images, and colourizing historical images.