Add ContentSegmentation in KnowledgeBase

[evan-cao-wb]

After the "Detection Model" generates a set of bounding boxes, certain post-processing steps become necessary to partition the entire image into distinct regions, laying the groundwork for the subsequent OCR stage.