简体中文| English
FastDeploy defines different structures (fastdeploy/vision/common/result.h) to express the model prediction results according to the vision model task.
| Structure | Document | Description | Corresponding Model |
|---|---|---|---|
| ClassifyResult | C++/Python document | Image classification return results | ResNet50, MobileNetV3, etc. |
| SegmentationResult | C++/Python document | Image segmentation result | PP-HumanSeg, PP-LiteSeg, etc. |
| DetectionResult | C++/Python document | Target detection result | PP-YOLOE, YOLOv7, etc. |
| FaceDetectionResult | C++/Python document | Result of face detection | SCRFD, RetinaFace, etc. |
| FaceAlignmentResult | C++/Python document | Face alignment result(Face keypoint detection) | PFLD model, etc. |
| KeyPointDetectionResult | C++/Python document | Result of keypoint detection | PP-Tinypose model, etc. |
| FaceRecognitionResult | C++/Python document | Result of face recognition | ArcFace, CosFace, etc. |
| MattingResult | C++/Python document | Image/video keying result | MODNet, RVM, etc. |
| OCRResult | C++/Python document | Text box detection, classification and text recognition result | OCR, etc. |
| MOTResult | C++/Python document | Multi-target tracking result | pptracking, etc. |
| HeadPoseResult | C++/Python document | Head pose estimation result | FSANet, etc. |