Segmentation by type delineates the methodological foundations that generate the labeled datasets feeding AI pipelines. Image Annotation (approximately 40% of service spend) includes sub‑types such as bounding box, polygon, semantic segmentation, and keypoint labeling, each offering varying granularity for computer‑vision models. Text Annotation (25%) covers entity recognition, sentiment tagging, and syntactic parsing, critical for natural‑language processing across industries. Video Annotation (15%) expands on frame‑by‑frame labeling with temporal bounding boxes, action segmentation, and object tracking, supporting surveillance and autonomous navigation. Audio Annotation (8%) involves speech‑to‑text transcription, speaker diarization, and sound event detection, enabling voice assistants and acoustic monitoring. Sensor Data Annotation (7%) addresses LiDAR, radar, and IoT streams through point‑cloud labeling and signal classification, essential for robotics and autonomous systems. 3D Point‑Cloud Annotation (5%) provides spatial context for AR/VR and autonomous platforms. Together, these types constitute the technical taxonomy that drives service pricing, labor intensity, and ultimately, the market’s revenue composition.