TI Deep Learning Product User Guide
|
This structure define the parameters of AnchorBox used by Detection Output Layer in TIDL.
int32_t sTIDL_AnchorBoxParams_t::numAnchors |
Num Anchors used by the Current Head
int32_t sTIDL_AnchorBoxParams_t::numKeyPoints |
Num Key Points to be detected, Default zero
int32_t sTIDL_AnchorBoxParams_t::headWidth |
Head Width as grid size
int32_t sTIDL_AnchorBoxParams_t::headHeight |
Head Height as grid size
float32_tidl sTIDL_AnchorBoxParams_t::strideW |
Normalized Stride in Horizontal Dir
float32_tidl sTIDL_AnchorBoxParams_t::strideH |
Normalized Stride in Vertical Dir
float32_tidl sTIDL_AnchorBoxParams_t::offsetW |
Normalized Offset in Horizontal Dir
float32_tidl sTIDL_AnchorBoxParams_t::offsetH |
Normalized Offset in Vertical Dir
float32_tidl sTIDL_AnchorBoxParams_t::offsetL |
Normalized Offset in Height Dir used in 3D OD
float32_tidl sTIDL_AnchorBoxParams_t::boxScales[4] |
Scales used for Box decoding Y,X,H,W
float32_tidl sTIDL_AnchorBoxParams_t::kpScales[TIDL_OD_MAX_KEYPOINTS] |
Scales used for Key points decoding
float32_tidl sTIDL_AnchorBoxParams_t::boxLength[TIDL_OD_MAX_ANCHORS_PER_HEAD] |
Normalized Anchors Box Width used in 3D OD only
float32_tidl sTIDL_AnchorBoxParams_t::boxWidth[TIDL_OD_MAX_ANCHORS_PER_HEAD] |
Normalized Anchors Box Width used in 3D and 2D OD
float32_tidl sTIDL_AnchorBoxParams_t::boxHeight[TIDL_OD_MAX_ANCHORS_PER_HEAD] |
Normalized Anchors Box height used in 3D and 2D OD
float32_tidl* sTIDL_AnchorBoxParams_t::anchorInputs |