--- title: "Physical AI Data and Robotics Annotation Services | YPAI" url: https://ypai.ai/annotation/robotics-industrial-vision-annotation-services/ description: "Annotation for robotic perception and industrial vision: taxonomy and guideline, edge-case rules, box, polygon, mask and keypoint labels, adjudication." source: "src/copy/routes (route /annotation/robotics-industrial-vision-annotation-services/)" --- # Easy frames are quick to label. The hard ones decide the model. > Annotation for robotic perception and industrial vision: taxonomy and guideline, edge-case rules, box, polygon, mask and keypoint labels, adjudication. Easy frames are quick to label. The hard ones decide the model. YPAI settles the taxonomy, the edge-case rules and the review method before production, so occlusion, thin boundaries, tiny parts and dense instances are labelled the same way in every frame. What the guideline has to settle ## Occlusion - how much of the object must be visible to count ## Thin boundary - where the edge sits when the contrast does not ## Tiny object - the size below which an instance is out of scope ## Dense instances - when touching objects are one label or many Drawing the shape is the easy half. Deciding what counts is the work. Two annotators can draw the same object and still disagree. What separates them is not skill, it is whether the guideline already answered the question: - what counts as the object at all - where the boundary sits when contrast fails - how much occlusion still earns a label - whether a truncated instance is kept or dropped - when touching objects are one instance or many - whether a deviation is a defect or in tolerance - which attributes travel with the instance - what an annotator does when the frame is ambiguous - who adjudicates, and what the record shows YPAI settles these questions in the guideline before production, not in review afterwards. The same holds for industrial vision: a surface mark becomes useful training data only when the guideline states the product, station, process stage, tolerance, viewpoint and inspection method that decide whether it is a defect at all. The taxonomy is built around the decision the system must make. - whether a surface contains a defect - whether the system should continue, stop, retry or escalate ## Start with the data bottleneck. YPAI can run one defined stage or the complete data operation. ## Write the guideline - Two annotators would answer the same frame differently. - A written guideline with worked examples, a frozen class list and the rules that resolve disagreement. ## Label a defined set - You have frames and an agreed acceptance criterion. - Labelled frames with per-instance attributes, reviewer records, agreement results and acceptance status. ## Re-label existing data - You have labels you cannot trust or cannot use across training and evaluation. - A re-labelled set against the current taxonomy, with the disagreements found and how each was resolved. ## Run the annotation operation - You need the work to keep running against a moving taxonomy. - A controlled labelling, review and acceptance workflow connected to your model-development environment. ## The edge is part of the task. A model trained on clean frames meets the other kind in production. These four decide the guideline, and the guideline decides whether two annotators agree. ### Occlusion - A forklift behind a pallet stack. How much of the object must remain visible before it stops being an instance. ### Thin boundary - Where the edge sits when the contrast does not. Wrapping, shadow and specular surfaces move the boundary a human would draw. ### Tiny object - Below some size an instance costs more to label than it teaches. That threshold is a decision, not a preference. ### Dense instances - Touching and stacked objects. One label or many, and whether the group is itself a class. Each one is settled in writing before production, with a worked example, so the answer is the same in frame one and frame forty thousand. ## Every class should say what it excludes. The exact taxonomy is written for the customer's product, station, model task and acceptance criteria. An annotation taxonomy can carry the following layers. ### Class definition ### Geometry ### Boundary rules ### Occlusion and truncation ### Instance separation ### Attributes ### Defect criteria ### Ambiguity and escalation - what the annotator does when unsure ### Review and acceptance The taxonomy is frozen before production scale. A class list is not a taxonomy. The rules that Every layer above is carried per instance. A guideline that reads well and cannot be applied to this frame is not finished. Each label type preserves something different and costs something different. The model task decides which one is right, not annotator preference. - Classification / one class per image, no location - Box / location and extent, fastest to review - Polygon / an irregular boundary, ordered vertices - Mask / a per-pixel region, highest cost to correct Keypoints and pose sit outside that progression: they preserve articulation rather than extent, and they are chosen when the model must act on the object, not only find it. ## Industrial vision data should represent the decision on the line. A defect label without process context may not be enough to train or evaluate the system that must act on it. The same surface mark is a defect at one station and in tolerance at the next. - The same warehouse frame under strong glare. - The same warehouse frame in low light. - The same warehouse frame with motion blur. - The same warehouse frame partly occluded. The delivered label set records not only the class and the geometry, but the capture conditions under which the inspection decision was made, so a disagreement can be traced to the frame rather than to the annotator. Rare classes and rare conditions are the ones a real capture rarely contains. Synthetic frames arrive pre-labelled, and they pass the same guideline, review and acceptance process as the real ones. Synthetic frames are a coverage tool, not a substitute for acceptance on real ones. ## Quality follows the complete label set. The quality process begins with the guideline, not with review, and ends only when the label-set version is accepted. Where accuracy or pass rate is the appropriate metric, YPAI typically proposes a 95% to 99% target range after calibration, with the exact metric, threshold and review method defined in the SOW. class list & taxonomy · annotation guideline with worked examples · reviewer and adjudication record · agreement report · quality report · provenance & rights record · label-set version record · checksums Delivered in or mapped to COCO, YOLO, Pascal VOC, KITTI, Parquet, JSONL or a customer-native schema. Format support is checked against the actual data, loader, toolchain and delivery requirement. Different systems need different data operations. We scope these separately: robot foundation models, manipulation and humanoid systems, industrial inspection, sensor and device programmes, and recovery of existing data. ## Send the images, the classes and the acceptance criteria. The system, the task, the data source, the sensors, the scale, the annotation and the delivery format. YPAI returns feasibility, method, schema, QA and acceptance, delivery format and the open decisions before launch. Scope a Physical AI data project Request a review of an existing dataset ## Related ### [Robotics at YPAI](https://ypai.ai/solutions/robotics/) - Collection, task demonstrations and held-out ground truth for robots in the environments they work in. ### [Data collection](https://ypai.ai/data-collection/) - Custom video, image, sensor and multimodal collection. ### [Video annotation](https://ypai.ai/annotation/video-annotation-services/) - Temporal events, tracking and action labels. ### [LiDAR and 3D](https://ypai.ai/annotation/lidar-3d-point-cloud-annotation-services/) ### [Sensor fusion](https://ypai.ai/annotation/sensor-fusion-annotation-services/) ### [Image annotation](https://ypai.ai/annotation/image-annotation/) ### [Managed annotation](https://ypai.ai/data-solutions/annotation/) ### [Pilots](https://ypai.ai/pilots/) - Validate the task, output and acceptance plan before scale.