Skip to content

Latest commit

 

History

History
110 lines (71 loc) · 8.39 KB

README.md

File metadata and controls

110 lines (71 loc) · 8.39 KB
Raw

Efficient Rotated NMS Plugin

Table of Contents

Description

This TensorRT plugin implements an efficient algorithm to perform Non Maximum Suppression for oriented bounding boxes object detection networks.

This plugin is primarily intended for using with oriented bounding boxes object detection networks on TensorRT, as this network is particularly sensitive to the latencies introduced by slower NMS implementations. However, the plugin is generic enough that it will work correctly for other detections architectures, such as YOLOv8-OBB.

This plugin is modified from TensorRT efficientNMSPlugin and MMDeploy TRTBatchedRotatedNMS .

Structure

Inputs

The plugin has two modes of operation, depending on the given input data. The plugin will automatically detect which mode to operate as, depending on the number of inputs it receives, as follows:

  1. Standard NMS Mode: Only two input tensors are given, (i) the bounding box coordinates and (ii) the corresponding classification scores for each box.

  2. Fused Box Decoder Mode: Three input tensors are given, (i) the raw localization predictions for each box originating directly from the localization head of the network, (ii) the corresponding classification scores originating from the classification head of the network, and (iii) the default anchor box coordinates usually hardcoded as constant tensors in the network.

Most object detection networks work by generating raw predictions from a "localization head" which adjust the coordinates of standard non-learned anchor coordinates to produce a tighter fitting bounding box. This process is called "box decoding", and it usually involves a large number of element-wise operations to transform the anchors to final box coordinates. As this can involve exponential operations on a large number of anchors, it can be computationally expensive, so this plugin gives the option of fusing the box decoder within the NMS operation which can be done in a far more efficient manner, resulting in lower latency for the network.

Boxes Input

Input Shape: [batch_size, number_boxes, 5] or [batch_size, number_boxes, number_classes, 5]

Data Type: float32 or float16

The boxes input can have 3 dimensions in case a single box prediction is produced for all classes (such as in EfficientDet or SSD), or 4 dimensions when separate box predictions are generated for each class (such as in FasterRCNN), in which case number_classes >= 1 and must match the number of classes in the scores input.

For Standard NMS Mode, this tensor should contain the final box coordinates for each predicted detection. For Fused Box Decoder Mode, this tensor should contain the raw localization predictions, i.e., the predicted box corrections relative to default anchor boxes. Specifically, the five values represent x1, y1, x2, y2, and r, or x, y, w, h, and r.

Scores Input

Input Shape: [batch_size, number_boxes, number_classes]

Data Type: float32 or float16

The scores input has number_classes elements with the predicted scores for each candidate class for each of the number_boxes anchor boxes.

Usually, the score values will have passed through a sigmoid activation function before reaching the NMS operation. However, as an optimization, the pre-sigmoid raw scores can also be provided to the NMS plugin to reduce overall network latency. If raw scores are given, enable the score_activation parameter so they are processed accordingly.

Anchors Input (Optional)

Input Shape: [1, number_boxes, 5] or [batch_size, number_boxes, 5]

Data Type: float32 or float16

Only used in Fused Box Decoder mode. It is much more efficient to perform the box decoding within this plugin. In this case, the boxes input will be treated as the raw localization head box corrections, and this third input should contain the default anchor/prior box coordinates.

When used, the input must have 3 dimensions, where the first one may be either 1 in case anchors are constant for all images in a batch, or batch_size in case each image has different anchors -- such as in the box refinement NMS of FasterRCNN's second stage.

Dynamic Shape Support

Most input shape dimensions, namely batch_size, number_boxes, and number_classes, for all inputs can be defined dynamically at runtime if the TensorRT engine is built with dynamic input shapes. However, once defined, these dimensions must match across all tensors that use them (e.g. the same number_boxes dimension must be given for both boxes and scores, etc.)

Box Coding Type

Different object detection networks represent their box coordinate system differently. The two types supported by this plugin are:

  1. RotatedBoxCorners: The four coordinates represent [x1, y1, x2, y2, r], where x1, y1 are the coordinates of the top-left corner of the bounding box, x2, y2 are the coordinates of the bottom-right corner, and r is the rotation angle of the box.

  2. RotatedBoxCenterSize: The four coordinates represent [x, y, w, h, r], where x, y define the center of the box, w, h define the width and height of the box, and r is the rotation angle of the box.

r represents the rotation angle of the bounding box and is used to describe the orientation of the box. For NMS purposes, horizontal and vertical coordinates are interchangeable, and the plugin will process these coordinates according to the input coding type.

In Fused Box Decoder mode, the boxes and anchor tensors should both use the same coding.

Outputs

The following four output tensors are generated:

  • num_detections: This is a [batch_size, 1] tensor of data type int32. The last dimension is a scalar indicating the number of valid detections per batch image. It can be less than max_output_boxes. Only the top num_detections[i] entries in nms_boxes[i], nms_scores[i] and nms_classes[i] are valid.

  • detection_boxes: This is a [batch_size, max_output_boxes, 5] tensor of data type float32 or float16, containing the coordinates of non-max suppressed boxes. The output coordinates will always be in BoxCorner format, regardless of the input code type.

  • detection_scores: This is a [batch_size, max_output_boxes] tensor of data type float32 or float16, containing the scores for the boxes.

  • detection_classes: This is a [batch_size, max_output_boxes] tensor of data type int32, containing the classes for the boxes.

Parameters

Type Parameter Description
float score_threshold * The scalar threshold for score (low scoring boxes are removed).
float iou_threshold The scalar threshold for IOU (additional boxes that have high IOU overlap with previously selected boxes are removed).
int max_output_boxes The maximum number of detections to output per image.
int background_class The label ID for the background class. If there is no background class, set it to -1.
bool score_activation * Set to true to apply sigmoid activation to the confidence scores during NMS operation.
bool class_agnostic Set to true to do class-independent NMS; otherwise, boxes of different classes would be considered separately during NMS.
int box_coding Coding type used for boxes (and anchors if applicable), 0 = BoxCorner, 1 = BoxCenterSize.

Parameters marked with a * have a non-negligible effect on runtime latency. See the Performance Tuning section below for more details on how to set them optimally.

Documentation