Use setBackend instead of ConvertToCpu workaround

Change the convertion to backend 'cpu' from a custom function converToCpu in the database to the generic setBackend().

Content :

• Remove the call to convertToCpu in database, its definition & tests.

However it raises a problem on the layout of the data between different backends. In the current code the resulting tensor with backend 'cpu' have layout [N,H,W,C] instead of [N,C,H,W].

• Use format HWC for OpenCV tensor and change format when converting to CPU backend.

include/aidge/backend/opencv/utils/Utils.hpp

@@ -33,27 +33,6 @@ namespace Aidge {
  */
 std::shared_ptr<Tensor> tensorOpencv(cv::Mat mat);
 
-/**
- * @brief Copy the data from a source 2D cv::mat to a destination pointer with an offset
- *
- * @tparam CV_T The standard type corresponding to the opencv data type
- * @param mat opencv 2D mat to copy the data from
- * @param data destination pointer
- * @param offset offset an the destination data pointer
- */
-template <class CV_T>
-void convert(const cv::Mat& mat, void* data, std::size_t offset);
-
-
-/**
- * @brief Convert a tensor backend opencv into a tensor backend cpu
- *
- * @param tensorOpencv tensor with backend opencv (contains a cv::mat)
- * @return std::shared_ptr<Tensor> tensor backend cpu (contains a std::vector)
- */
-std::shared_ptr<Tensor> convertCpu(std::shared_ptr<Aidge::Tensor> tensorOpencv);
-
-
 }  // namespace
 
 #endif // AIDGE_OPENCV_UTILS_UTILS_H_
\ No newline at end of file