Scientific Chart image property identification by connected component labeling in PDF files

Scientific Chart Recognition (SCR) in documents is an area of research work that has gained attention only in the past few decades. This paper proposes the use of single pass Connected Component Labels (CCL) to separate the graphics regions from the text region. A box bounding algorithm is then used to locate the scientific chart from the graphics region. The experimental result shows that the accuracy and processing speed is improved by the introduction of the single pass CCL.