DocumentFigureClassifier-v2.5 is an image classification model based on EfficientNet-B0, fine-tuned to recognize 26 different types of figures commonly found in documents. It can distinguish between charts, tables, photographs, diagrams, signatures, and many other visual element types. The model is provided in multiple formats including ONNX and is designed to be integrated into document analysis pipelines.
DocumentFigureClassifier is an Other AI product. Automatically identifying and categorizing visual elements such as charts, diagrams, and photos within PDF documents. DocumentFigureClassifier is an open-source project aimed at document processing developers. The project is open source (Open Source). DocumentFigureClassifier is available on the web and API.
It is developed by Docling, and the product first shipped in 2024. Among its 4 catalogued features are Image Classification, Document Analysis, and 26 Figure Categories.
Latest indexed changes and source events
docling-project/DocumentFigureClassifier-v2.5 verified by the PulseGate indexer
Other apps tracked under the same category.
