opencv/modules/objdetect/include
quantizs e1b06371ad Added buffer-based model loading to FaceRecognizerSF
- Implemented a new `create` method in `FaceRecognizerSF` to allow model and configuration loading from memory buffers (std::vector<uchar>), similar to the existing functionality in `FaceDetectorYN`.
- Updated `face_recognize.cpp` with a new constructor in `FaceRecognizerSFImpl` that supports buffer-based loading for both model weights and network configuration.
- Ensured compatibility with both file-based and buffer-based model loading by maintaining consistent backend and target settings across both constructors.
- This change improves flexibility, allowing FaceRecognizerSF to be instantiated from memory buffers, which is useful for dynamic model loading scenarios such as embedded systems or applications where models are loaded in-memory.
2024-10-09 15:13:47 +02:00
..
opencv2 Added buffer-based model loading to FaceRecognizerSF 2024-10-09 15:13:47 +02:00