Tech Academy
The core 2D/3D vision, AI model and Physical AI technology inside finetheAi solutions, explained step by step. Ready to use for customer training and onboarding.
2D/3D Vision
Acquiring, aligning and measuring images — the step that sets inspection quality before any AI runs.
- 01
Good Images Come First: Camera, Lighting, Trigger and Preprocessing
An AI model can see only what is in the image. This lesson explains how to choose the camera, lighting and capture timing so defects stay clear in the image, and how to clean up brightness and background before the image goes to the model.
- 02
Alignment: Putting Every Part in the Same Pose
If parts sit at a different angle every time, the model learns rotation before it learns defects. This lesson covers the principles and uses of alignment: Ring alignment for round parts, PCA alignment for arbitrary shapes, and yolo_align, which judges after alignment.
- 03
Rule-Based Dimensional Measurement: Turning Pixels into Millimeters
For inspections where the tolerance is a number, measurement fits better than AI. This lesson explains how to measure diameter, thickness and roundness with calibration, contour extraction and circle fitting, and how to combine them with AI judgment.
- 04
Photometric Stereo: Seeing Surface Relief with a 2D Camera
If you capture several images while changing the lighting direction, you can compute which way the surface faces at each pixel. This lesson explains how a curvature map reveals shallow dents that the eye and ordinary images cannot show.
- 05
3D Vision and Point Clouds: When You Need Height and Pose
A 2D image is not enough for a robot to pick a part or to measure weld bead height. This lesson explains how to clean up, split and align the point cloud from a 3D camera.
AI Models
Anomaly detection learned from good parts only, object detection that names and locates defects, and how to evaluate and run them on the line.
- 01
Finding Anomalies from Normal Images Only: Anomaly Detection and PatchCore
Defect samples are hard to collect, and tomorrow brings a defect you have never seen. This lesson explains anomaly detection, which trains on normal images only and finds what differs from normal, using the widely used PatchCore as the example.
- 02
Choosing an Anomaly Detection Algorithm: 11 Options and When to Use Each
The 11 algorithms in AD Workbench, including PaDiM, PatchCore, AnomalyDINO, WinCLIP and EfficientAD, produce the same outputs but have different strengths. This lesson gives criteria for choosing by data volume, part fixturing, resolution and takt time.
- 03
Object Detection: YOLO and RT-DETR Give Defects a Name and a Location
This lesson explains object detection, which answers what is there, where, and how sure the model is, for defect types you already know. It covers box, class and confidence, IoU and NMS, per-class thresholds, and how to use it with anomaly detection.
- 04
Model Evaluation and Field Operations: False Rejects, Escapes, Thresholds, Retraining
A 99% accuracy number alone is not enough to put a model on the line. This lesson explains how to count false rejects and escapes separately, set the threshold to match cost, and monitor changes in operation so you can retrain.
Physical AI
Robots acting on what vision sees — coordinate calibration, digital twins, and harvesting and picking automation.
- 01
Physical AI: AI That Sees, Decides and Moves
Physical AI goes beyond showing inspection results on a screen. The robot uses those results to pick, move and finish objects. This lesson explains the components and the operating loop with a shiitake mushroom harvesting case.
- 02
Hand-Eye Calibration: Sending the Robot Where the Camera Sees
The camera says 'it is here' in its own coordinates, and the robot moves only in its own coordinates. This lesson explains the concept, procedure and accuracy tips of Hand-Eye calibration, which finds the transform between them.
- 03
Digital Twin and Pick & Place: Verify Virtually Before the Real Robot Moves
If you copy a robot cell into virtual space, you can check paths, collisions and cycle time before the real robot moves. This lesson follows the whole Pick & Place process, from 3D recognition to grasp planning and live mirroring.