Caliper recognizes similar industrial parts using images and size data
CALIPER: Metric-Grounded Model-Free Recognition of Visually Similar Industrial Parts
Computer Vision and Pattern Recognition
Summary
Recognizing industrial parts that look very similar is tough because they mainly differ in size, which normal image processing can miss. The authors propose CALIPER, a method that uses both visual appearance from RGB-D images and exact size information to tell parts apart. It learns from just a few example images without needing big datasets or CAD models, making it easy to add new parts. Their tests show it accurately identifies parts and can work on a robot arm without retraining.
What this means in practice
- •For industrial automation teams: Identify visually similar parts on assembly lines using limited data and without needing CAD models for each part.
- •For robotics system integrators: Deploy parts recognition systems on robot arms that generalize to new parts without retraining the entire network.
Authors
Alankrit Gupta, Chenxi Tao, Seung-Kyum Choi
Abstract
Fine-grained recognition of visually similar industrial parts is challenging when classes differ primarily in physical dimensions. Normalizing detected object crops to a fixed input size suppresses absolute scale, while CAD models and large class-specific datasets may be unavailable in evolving industrial inventories. We present CALIPER, a model-free RGB-D framework that couples support-based appearance matching with metric size evidence. Each training class is onboarded from a single turntable RGB-D video and one to two labeled real images; 3D reconstruction provides novel-view appearance support, while aligned depth yields a class-specific metric size profile. At inference, a coarse YOLOv8n-seg model localizes parts, and a frozen DINOv2 backbone with an episodically trained embedding head performs fine-grained support matching. Margin-conditioned metric fusion activates probabilistic size evidence only for appearance-ambiguous decisions. New classes are enrolled from a small RGB-D support set without updating network parameters. We evaluate CALIPER on 18 visually similar industrial parts: 16 classes are used for training, while two screws are reserved for training-free enrollment. CALIPER achieves 88.2% closed-set accuracy with 99.8% localization recall and 85.7% overall accuracy after 10-shot enrollment of the two unseen screws. Metric fusion improves unseen-class accuracy by up to 37.4 percentage points without statistically significant degradation of the original inventory. Robot-arm deployment identifies 17/18 parts without deployment-specific retraining.