NonMaxSuppression

Versioned name: NonMaxSuppression-5

Category: Sorting and maximization

Short description: NonMaxSuppression performs non maximum suppression of the boxes with predicted scores.

Detailed description: NonMaxSuppression performs non maximum suppression algorithm as described below:

1. Let B = [b_0,...,b_n] be the list of initial detection boxes, S = [s_0,...,s_N] be the list of corresponding scores.
2. Let D = [] be an initial collection of resulting boxes.
3. If B is empty then go to step 8.
4. Take the box with highest score. Suppose that it is the box b with the score s.
5. Delete b from B.
6. If the score s is greater or equal than score_threshold then add b to D else go to step 8.
7. For each input box b_i from B and the corresponding score s_i, set s_i = s_i * func(IOU(b_i, b)) and go to step 3.
8. Return D, a collection of the corresponding scores S, and the number of elements in D.

Here func(iou) = 1 if iou <= iou_threshold else 0 when soft_nms_sigma == 0, else func(iou) = exp(-0.5 * iou * iou / soft_nms_sigma) if iou <= iou_threshold else 0.

This algorithm is applied independently to each class of each batch element. The total number of output boxes for each class must not exceed max_output_boxes_per_class.

Attributes:

• box_encoding
  • Description: box_encoding specifies the format of boxes data encoding.
  • Range of values: "corner" or "center"
    • corner - the box data is supplied as [y1, x1, y2, x2] where (y1, x1) and (y2, x2) are the coordinates of any diagonal pair of box corners.
    • center - the box data is supplied as [x_center, y_center, width, height].
  • Type: string
  • Default value: "corner"
  • Required: no
• sort_result_descending
  • Description: sort_result_descending is a flag that specifies whenever it is necessary to sort selected boxes across batches or not.
  • Range of values: true of false
    • true - sort selected boxes across batches.
    • false - do not sort selected boxes across batches (boxes are sorted per class).
  • Type: boolean
  • Default value: true
  • Required: no
• output_type
  • Description: the output tensor type
  • Range of values: "i64" or "i32"
  • Type: string
  • Default value: "i64"
  • Required: No

Inputs:

• 1: boxes - tensor of type T and shape [num_batches, num_boxes, 4] with box coordinates. Required.
• 2: scores - tensor of type T and shape [num_batches, num_classes, num_boxes] with box scores. Required.
• 3: max_output_boxes_per_class - scalar or 1D tensor with 1 element of type T_MAX_BOXES specifying maximum number of boxes to be selected per class. Optional with default value 0 meaning select no boxes.
• 4: iou_threshold - scalar or 1D tensor with 1 element of type T_THRESHOLDS specifying intersection over union threshold. Optional with default value 0 meaning keep all boxes.
• 5: score_threshold - scalar or 1D tensor with 1 element of type T_THRESHOLDS specifying minimum score to consider box for the processing. Optional with default value 0.
• 6: soft_nms_sigma - scalar or 1D tensor with 1 element of type T_THRESHOLDS specifying the sigma parameter for Soft-NMS; see Bodla et al. Optional with default value 0.

Outputs:

• 1: selected_indices - tensor of type T_IND and shape [number of selected boxes, 3] containing information about selected boxes as triplets [batch_index, class_index, box_index].
• 2: selected_scores - tensor of type T_THRESHOLDS and shape [number of selected boxes, 3] containing information about scores for each selected box as triplets [batch_index, class_index, box_score].
• 3: valid_outputs - 1D tensor with 1 element of type T_IND representing the total number of selected boxes.

Plugins which do not support dynamic output tensors produce selected_indices and selected_scores tensors of shape [min(num_boxes, max_output_boxes_per_class) * num_batches * num_classes, 3] which is an upper bound for the number of possible selected boxes. Output tensor elements following the really selected boxes are filled with value -1.

Types

• T: floating point type.
• T_MAX_BOXES: integer type.
• T_THRESHOLDS: floating point type.
• T_IND: int64 or int32.

Example

<layer ... type="NonMaxSuppression" ... >
    <data box_encoding="corner" sort_result_descending="1" output_type="i64"/>
    <input>
        <port id="0">
            <dim>3</dim>
            <dim>100</dim>
            <dim>4</dim>
        </port>
        <port id="1">
            <dim>3</dim>
            <dim>5</dim>
            <dim>100</dim>
        </port>
        <port id="2"/> <!-- 10 -->
        <port id="3"/>
        <port id="4"/>
    </input>
    <output>
        <port id="5" precision="I64">
            <dim>150</dim> <!-- min(100, 10) * 3 * 5 -->
            <dim>3</dim>
        </port>
        <port id="6" precision="FP32">
            <dim>150</dim> <!-- min(100, 10) * 3 * 5 -->
            <dim>3</dim>
        </port>
        <port id="7" precision="I64">
            <dim>1</dim>
        </port>
    </output>
</layer>