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. |