Human Related Information Extraction from Chinese Archive Images

With the rise of the information age, digitalization and paperless processes have become the norm in managing archive images. However, research on extracting and managing image content from archives is still in its early stages, and is primarily focused on recognizing fixed-format archive images. As a result, there is a lack of technology for extracting key personal information applicable to all types of archive images. To address this, we have identified two main tasks: extracting identity photos and key personal information. To ensure confidentiality of real data, we created a dataset that simulates certificate photo files. We then used a YOLOv5-based object detection network to train a model that can detect document photos in archive images. We also used a combination of PP-OCR text recognition and object detection to extract key information from document images.