Photo Album Creation for Historical Newspaper through Computer Vision by Using ABBYY, VGG Model, and Yolov5
Vincent Y. Lum, Kin-Lok Yeung Calvin · 2023
In this paper we presented a novel approach with Computer Vision and Machine/Deep Learning for building digital image albums and storytelling with images from historical newspapers. In total 456 newspaper pages have been used to execute the proposed method. Different methodologies, including ABBYY FineReader, VGG nls-chapbook model, and Yolov5, were compared and evaluated. Image categorization has also been studied with Face Recognition and OCR methodologies. Experimental results showed that the proposed method successfully extracted and categorized the illustrations of 106 actors/actresses which were then used to create their personal photo albums. Our experiment with image detection for photo album building demonstrated the potential of our approach to fit other newspaper or digitized collections. The result photo album of historical personalities allowed users to have a more comprehensive understanding of the newspaper article mentioned about the actors/actresses and their multifaceted biographical information as well.