RoiAlign - 10 vs 22¶
Next section compares an older to a newer version of the same operator after both definition are converted into markdown text. Green means an addition to the newer version, red means a deletion. Anything else is unchanged.
- RoiAlign10 → RoiAlign22 +5 -1
RoiAlign10 → RoiAlign22
RENAMED
|
@@ -1 +1 @@
|
|
|
1
1
|
Region of Interest (RoI) align operation described in the
|
|
2
2
|
[Mask R-CNN paper](https://arxiv.org/abs/1703.06870).
|
|
3
3
|
RoiAlign consumes an input tensor X and region of interests (rois)
|
|
4
4
|
to apply pooling across each RoI; it produces a 4-D tensor of shape
|
|
5
5
|
(num_rois, C, output_height, output_width).
|
|
6
6
|
RoiAlign is proposed to avoid the misalignment by removing
|
|
7
7
|
quantizations while converting from original image into feature
|
|
8
8
|
map and from feature map into RoI feature; in each ROI bin,
|
|
9
9
|
the value of the sampled locations are computed directly
|
|
10
10
|
through bilinear interpolation.
|
|
11
11
|
### Attributes
|
|
12
|
+
|
|
13
|
+
* **coordinate_transformation_mode - STRING** (default is 'half_pixel'):
|
|
14
|
+
|
|
15
|
+
Allowed values are 'half_pixel' and 'output_half_pixel'. Use the value 'half_pixel' to pixel shift the input coordinates by -0.5 (the recommended behavior). Use the value 'output_half_pixel' to omit the pixel shift for the input (use this for a backward-compatible behavior).
|
|
12
16
|
* **mode - STRING** (default is 'avg'):
|
|
13
17
|
The pooling method. Two modes are supported: 'avg' and 'max'. Default is 'avg'.
|
|
14
18
|
* **output_height - INT** (default is '1'):
|
|
15
19
|
default 1; Pooled output Y's height.
|
|
16
20
|
* **output_width - INT** (default is '1'):
|
|
17
21
|
default 1; Pooled output Y's width.
|
|
18
22
|
* **sampling_ratio - INT** (default is '0'):
|
|
19
23
|
Number of sampling points in the interpolation grid used to compute the output value of each pooled output bin. If > 0, then exactly sampling_ratio x sampling_ratio grid points are used. If == 0, then an adaptive number of grid points are used (computed as ceil(roi_width / output_width), and likewise for height). Default is 0.
|
|
20
24
|
* **spatial_scale - FLOAT** (default is '1.0'):
|
|
21
25
|
Multiplicative spatial scale factor to translate ROI coordinates from their input spatial scale to the scale used when pooling, i.e., spatial scale of the input feature map X relative to the input image. E.g.; default is 1.0f.
|
|
22
26
|
### Inputs
|
|
23
27
|
- **X** (heterogeneous) - **T1**:
|
|
24
28
|
Input data tensor from the previous operator; 4-D feature map of shape (N, C, H, W), where N is the batch size, C is the number of channels, and H and W are the height and the width of the data.
|
|
25
29
|
- **rois** (heterogeneous) - **T1**:
|
|
26
30
|
RoIs (Regions of Interest) to pool over; rois is 2-D input of shape (num_rois, 4) given as [[x1, y1, x2, y2], ...]. The RoIs' coordinates are in the coordinate system of the input image. Each coordinate set has a 1:1 correspondence with the 'batch_indices' input.
|
|
27
31
|
- **batch_indices** (heterogeneous) - **T2**:
|
|
28
32
|
1-D tensor of shape (num_rois,) with each element denoting the index of the corresponding image in the batch.
|
|
29
33
|
### Outputs
|
|
30
34
|
- **Y** (heterogeneous) - **T1**:
|
|
31
35
|
RoI pooled output, 4-D tensor of shape (num_rois, C, output_height, output_width). The r-th batch element Y[r-1] is a pooled feature map corresponding to the r-th RoI X[r-1].
|
|
32
36
|
### Type Constraints
|
|
33
|
-
* **T1** in ( tensor(double), tensor(float), tensor(float16) ):
|
|
37
|
+
* **T1** in ( tensor(bfloat16), tensor(double), tensor(float), tensor(float16) ):
|
|
34
38
|
Constrain types to float tensors.
|
|
35
39
|
* **T2** in ( tensor(int64) ):
|
|
36
40
|
Constrain types to int tensors.
|