This op produces Region of Interests from given bounding boxes(bbox_deltas) encoded wrt anchors according to eq.2 in arXiv:1506.01497
The op selects top `pre_nms_topn` scoring boxes, decodes them with respect to anchors, applies non-maximal suppression on overlapping boxes with higher than `nms_threshold` intersection-over-union (iou) value, discarding boxes where shorter side is less than `min_size`. Inputs: `scores`: A 4D tensor of shape [Batch, Height, Width, Num Anchors] containing the scores per anchor at given position `bbox_deltas`: is a tensor of shape [Batch, Height, Width, 4 x Num Anchors] boxes encoded to each anchor `anchors`: A 1D tensor of shape [4 x Num Anchors], representing the anchors. Outputs: `rois`: output RoIs, a 3D tensor of shape [Batch, post_nms_topn, 4], padded by 0 if less than post_nms_topn candidates found. `roi_probabilities`: probability scores of each roi in 'rois', a 2D tensor of shape [Batch,post_nms_topn], padded with 0 if needed, sorted by scores.
Nested Classes
class | GenerateBoundingBoxProposals.Options |
Optional attributes for
GenerateBoundingBoxProposals
|
Public Methods
static GenerateBoundingBoxProposals |
create
(
Scope
scope,
Operand
<Float> scores,
Operand
<Float> bboxDeltas,
Operand
<Float> imageInfo,
Operand
<Float> anchors,
Operand
<Float> nmsThreshold,
Operand
<Integer> preNmsTopn,
Operand
<Float> minSize,
Options...
options)
Factory method to create a class wrapping a new GenerateBoundingBoxProposals operation.
|
static GenerateBoundingBoxProposals.Options |
postNmsTopn
(Long postNmsTopn)
|
Output <Float> |
roiProbabilities
()
A 2-D float tensor of shape `[num_images, post_nms_topn]` representing the score of the
region of interest box in `rois` tensor at the same index.
|
Output <Float> |
rois
()
A 3-D float tensor of shape `[num_images,post_nms_topn,4]` representing the selected
region of interest boxes.
|
Inherited Methods
Public Methods
public static GenerateBoundingBoxProposals create ( Scope scope, Operand <Float> scores, Operand <Float> bboxDeltas, Operand <Float> imageInfo, Operand <Float> anchors, Operand <Float> nmsThreshold, Operand <Integer> preNmsTopn, Operand <Float> minSize, Options... options)
Factory method to create a class wrapping a new GenerateBoundingBoxProposals operation.
Parameters
scope | current scope |
---|---|
scores | A 4-D float tensor of shape `[num_images, height, width, num_achors]` containing scores of the boxes for given anchors, can be unsorted. |
bboxDeltas | A 4-D float tensor of shape `[num_images, height, width, 4 x num_anchors]`. encoding boxes with respec to each anchor. Coordinates are given in the form [dy, dx, dh, dw]. |
imageInfo | A 2-D float tensor of shape `[num_images, 5]` containing image information Height, Width, Scale. |
anchors | A 2-D float tensor of shape `[num_anchors, 4]` describing the anchor boxes. Boxes are formatted in the form [y1, x1, y2, x2]. |
nmsThreshold | A scalar float tensor for non-maximal-suppression threshold. |
preNmsTopn | A scalar int tensor for the number of top scoring boxes to be used as input. |
minSize | A scalar float tensor. Any box that has a smaller size than min_size will be discarded. |
options | carries optional attributes values |
Returns
- a new instance of GenerateBoundingBoxProposals
public static GenerateBoundingBoxProposals.Options postNmsTopn (Long postNmsTopn)
Parameters
postNmsTopn | An integer. Maximum number of rois in the output. |
---|
public Output <Float> roiProbabilities ()
A 2-D float tensor of shape `[num_images, post_nms_topn]` representing the score of the region of interest box in `rois` tensor at the same index.
public Output <Float> rois ()
A 3-D float tensor of shape `[num_images,post_nms_topn,4]` representing the selected region of interest boxes. Sorted in descending order in scores.