Blueprint Studio User Experience Improvement Program

3D Mesh Geometric Feature Data: Only the geometric and mathematical features of 3D meshes are extracted; the 3D mesh models themselves are not extracted. This includes: the topological structure of the 3D mesh, the number of triangular facets, bounding box dimensions, curvature features, and the geometric parameters of the lowest point and overhang regions.

Slicing Process Configuration and Decision Data: Output labels of the intelligent model classification algorithm, the rotation angle and spatial placement matrix of the workpiece on the build platform, generated support point coordinates, support pillar diameters and base contact surface parameters, layer thickness, and exposure process parameters.

Device Print Result Feedback (collected via the desktop client): Physical print success/failure status reported through the print scheduling system (e.g., classification of phenomena such as print success, overhang collapse, local delamination failure, and part breakage), consumable material type, build environment temperature, and actual print duration.

Runtime Environment Diagnostic Logs: Blueprint Studio client version, ONNX inference engine latency, graphics/compute hardware acceleration environment, and abnormal crash error codes.

Explicitly Excluded Data: This Program never collects any personally identifiable information (including real names, identification documents, contact information, and payment accounts), account passwords, non-additive-manufacturing engineering files stored locally on the computer, or personal privacy metadata in model files that is unrelated to geometric structure.

Optimize the 3D Mesh Classification Algorithm: Based on the diverse heterogeneous geometric models imported from real production, expand the training sample set of "pose pre-alignment → feature extraction → clustering" to improve the hit rate and generalization capability for classifying long-tail and unconventional model shapes.

Fine-tune the Automatic Mesh Placement Algorithm (one class, one model): Combined with physical print success/failure cases reported through closed-loop feedback from the desktop client, incrementally train the independent placement model for specific geometric classes to reduce abrupt projection area changes and cup-shaped negative pressure effects, and predict the ideal posture that minimizes peel force.

Strengthen the Intelligent Support Generation Algorithm: For workpieces affected by breakage, surface damage caused by over-support, or supports that fail without bearing load, perform reinforcement learning training on the baseline model and the fine-tuning set to dynamically optimize contact point penetration depth, support stiffness, and overhang reinforcement strategies.

Drive Cloud-Based Incremental Hot Updates of Models: After benchmark evaluation of the accuracy of the fine-tuning set and test set is completed in the cloud, HeyGears Technology distributes the qualified ONNX model files to clients, enabling selective hot updates of partial algorithms without requiring a major software version upgrade.

De-identification and Desensitization: Before upload, all user identity identifiers are stripped locally from 3D meshes and slicing engineering data; only anonymous engineering task IDs generated by a random hash algorithm are assigned, which cannot be traced back to a specific organization or individual.

End-to-End Encrypted Transmission and Storage: All transmission links are mandatorily protected with standard TLS/HTTPS encryption, and the cloud training cluster uses AES-256 storage encryption and network-level physical isolation.

Dedicated Use and Access Control: Data may only be accessed by HeyGears Technology's algorithm R&D and slicing process teams within a restricted computing environment; it is strictly prohibited to use it for any commercial advertising or marketing, transfer it to third parties, or use it for unauthorized purposes.

Voluntary Participation and Opt-Out at Any Time: You may decide independently, during initial installation and configuration or in subsequent use, whether to enable this Program. You can check or uncheck "Join the User Experience Improvement Program" at any time in the client under Settings → General → Data & Privacy.

Effect of Opting Out: Once you uncheck the option, Blueprint Studio will immediately stop uploading any slicing data and device success/failure feedback. Opting out of this Program does not affect your normal use of slicing and layout, manual support addition, offline slicing, or the existing offline inference features.