MSL

MSL extracts and analyzes text and images from PDF-formatted scientific literature to retrieve experimental and biomedical figure data for downstream analysis.

Key Features:

• Advanced data mining and image processing: Employs data mining and image processing techniques to extract and interpret text and images from PDF files.
• Text marginalization: Supports marginalization of extracted text based on spatial coordinates and keyword filters.
• Figure visualization and OCR-based text extraction: Visualizes extracted figures and applies Optical Character Recognition (OCR) to recover embedded text from biological and biomedical figures.
• Physical and logical document analysis: Uses both physical and logical analysis of documents to locate and segment textual and graphical elements within PDFs.
• Multi-format output generation: Produces outputs in plain text, PDF, XML, and image file formats.

Scientific Applications:

• Extraction of experimental figure data: Retrieves data from experimental modalities including PCR-ELISA, microarray analysis, gel electrophoresis, and mass spectrometry.
• Sequencing and molecular data access: Extracts figures and associated text presenting DNA/RNA sequencing results.
• Diagnostic and medicinal imaging retrieval: Extracts and analyzes published imaging data such as CT, MRI, ultrasound, EEG, MEG, ECG, and PET figures.
• Support for secondary analyses: Facilitates replication, meta-analyses, and systematic reviews by providing access to original figure-derived data.

Methodology:

Implemented on a product line architecture using physical and logical document analysis, data mining and image processing, OCR for embedded text extraction, text marginalization by coordinates/keywords, figure visualization, and export to plain text, PDF, XML, and image files.