YOLOv12 with TensorRT backend

Author: Nuzhny007

README.md

@@ -3,6 +3,8 @@
 
 # Last changes
 
+* YOLOv12 detector worked with TensorRT! Export pretrained Pytorch models [here (sunsmarterjie/yolov12)](https://github.com/sunsmarterjie/yolov12) to onnx format and run Multitarget-tracker with -e=6 example
+
 * TensorRT 10 is supported
 
 * YOLOv11, YOLOv11-obb and YOLOv11-seg detector worked with TensorRT! Export pretrained Pytorch models [here (ultralytics/ultralytics)](https://github.com/ultralytics/ultralytics) to onnx format and run Multitarget-tracker with -e=6 example

data/settings_yolov12.ini

@@ -0,0 +1,142 @@
+[detection]
+
+#-----------------------------
+# opencv_dnn = 12
+# darknet_cudnn = 10
+# tensorrt = 11
+detector_backend = 11
+
+#-----------------------------
+# Target and backend for opencv_dnn detector
+# DNN_TARGET_CPU
+# DNN_TARGET_OPENCL
+# DNN_TARGET_OPENCL_FP16
+# DNN_TARGET_MYRIAD
+# DNN_TARGET_CUDA
+# DNN_TARGET_CUDA_FP16
+ocv_dnn_target = DNN_TARGET_CPU
+
+# DNN_BACKEND_DEFAULT
+# DNN_BACKEND_HALIDE
+# DNN_BACKEND_INFERENCE_ENGINE
+# DNN_BACKEND_OPENCV
+# DNN_BACKEND_VKCOM
+# DNN_BACKEND_CUDA
+# DNN_BACKEND_INFERENCE_ENGINE_NGRAPH
+# DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019
+ocv_dnn_backend = DNN_BACKEND_OPENCV
+
+#-----------------------------
+nn_weights = C:/work/home/mtracker/Multitarget-tracker/data/coco/yolov12m.onnx
+nn_config = C:/work/home/mtracker/Multitarget-tracker/data/coco/yolov12m.onnx
+class_names = C:/work/home/mtracker/Multitarget-tracker/data/coco/coco.names
+
+#-----------------------------
+confidence_threshold = 0.5
+	
+max_crop_ratio = 0
+max_batch = 1
+gpu_id = 0
+
+#-----------------------------
+# YOLOV3 
+# YOLOV4 
+# YOLOV5 
+net_type = YOLOV11
+
+#-----------------------------
+# INT8
+# FP16
+# FP32
+inference_precision = FP16
+
+
+[tracking]
+
+#-----------------------------
+# DistCenters = 0   // Euclidean distance between centers, pixels
+# DistRects = 1     // Euclidean distance between bounding rectangles, pixels
+# DistJaccard = 2   // Intersection over Union, IoU, [0, 1]
+# DistHist = 3      // Bhatacharia distance between histograms, [0, 1]
+
+distance_type = 0
+
+#-----------------------------
+# KalmanLinear = 0
+# KalmanUnscented = 1
+
+kalman_type = 0
+
+#-----------------------------
+# FilterCenter = 0
+# FilterRect = 1
+# FilterRRect = 2
+
+filter_goal = 0
+
+#-----------------------------
+# TrackNone = 0
+# TrackKCF = 1
+# TrackMIL = 2
+# TrackMedianFlow = 3
+# TrackGOTURN = 4
+# TrackMOSSE = 5
+# TrackCSRT = 6
+# TrackDAT = 7
+# TrackSTAPLE = 8
+# TrackLDES = 9
+# TrackDaSiamRPN = 10
+# Used if filter_goal == FilterRect
+
+lost_track_type = 0
+
+#-----------------------------
+# MatchHungrian = 0
+# MatchBipart = 1
+
+match_type = 0
+
+#-----------------------------
+# Use constant acceleration motion model:
+# 0 - unused (stable)
+# 1 - use acceleration in Kalman filter (experimental)
+use_aceleration = 0
+
+#-----------------------------
+# Delta time for Kalman filter
+delta_time = 0.4
+
+#-----------------------------
+# Accel noise magnitude for Kalman filter
+accel_noise = 0.2
+
+#-----------------------------
+# Distance threshold between region and object on two frames
+dist_thresh = 0.8 
+
+#-----------------------------
+# If this value > 0 than will be used circle with this radius
+# If this value <= 0 than will be used ellipse with size (3*vx, 3*vy), vx and vy - horizontal and vertical speed in pixelsa
+min_area_radius_pix = -1
+
+#-----------------------------
+# Minimal area radius in ration for object size. Used if min_area_radius_pix < 0
+min_area_radius_k = 0.8
+
+#-----------------------------
+# If the object do not assignment more than this frames then it will be removed
+max_skip_frames = 50
+
+#-----------------------------
+# The maximum trajectory length
+max_trace_len = 50
+
+#-----------------------------
+# Detection abandoned objects
+detect_abandoned = 0
+# After this time (in seconds) the object is considered abandoned
+min_static_time = 5
+# After this time (in seconds) the abandoned object will be removed
+max_static_time = 25
+# Speed in pixels. If speed of object is more that this value than object is non static
+max_speed_for_static = 10

example/examples.h

@@ -655,7 +655,8 @@ class YoloTensorRTExample final : public VideoExample
 				YOLOv10,
 				YOLOv11,
 				YOLOv11_OBB,
-				YOLOv11Mask
+				YOLOv11Mask,
+				YOLOv12
             };
             YOLOModels usedModel = YOLOModels::YOLOv9;
             switch (usedModel)

src/Detector/OCVDNNDetector.cpp

@@ -167,6 +167,7 @@ bool OCVDNNDetector::Init(const config_t& config)
         dictNetType["YOLOV11"] = ModelType::YOLOV11;
         dictNetType["YOLOV11_OBB"] = ModelType::YOLOV11_OBB;
         dictNetType["YOLOV11Mask"] = ModelType::YOLOV11Mask;
+        dictNetType["YOLOV12"] = ModelType::YOLOV12;
 
         auto netType = dictNetType.find(net_type->second);
         if (netType != dictNetType.end())
@@ -333,6 +334,9 @@ void OCVDNNDetector::DetectInCrop(const cv::UMat& colorFrame, const cv::Rect& cr
 	case ModelType::YOLOV11:
         ParseYOLOv11(crop, detections, tmpRegions);
         break;
+    case ModelType::YOLOV12:
+        ParseYOLOv11(crop, detections, tmpRegions);
+        break;
 
     case ModelType::YOLOV5_OBB:
     case ModelType::YOLOV8_OBB:

src/Detector/OCVDNNDetector.h

@@ -47,7 +47,8 @@ class OCVDNNDetector final : public BaseDetector
         YOLOV10,
         YOLOV11,
         YOLOV11_OBB,
-        YOLOV11Mask
+        YOLOV11Mask,
+        YOLOV12
     };
 
     cv::dnn::Net m_net;

src/Detector/tensorrt_yolo/YoloONNXv10_bb.hpp

@@ -7,6 +7,13 @@
 ///
 class YOLOv10_bb_onnx : public YoloONNX
 {
+public:
+	YOLOv10_bb_onnx(std::vector<std::string>& inputTensorNames, std::vector<std::string>& outputTensorNames)
+	{
+		inputTensorNames.push_back("images");
+		outputTensorNames.push_back("output0");
+	}
+
 protected:
 	///
 	/// \brief GetResult

src/Detector/tensorrt_yolo/YoloONNXv11_bb.hpp

@@ -7,6 +7,13 @@
 ///
 class YOLOv11_bb_onnx : public YoloONNX
 {
+public:
+	YOLOv11_bb_onnx(std::vector<std::string>& inputTensorNames, std::vector<std::string>& outputTensorNames)
+	{
+		inputTensorNames.push_back("images");
+		outputTensorNames.push_back("output0");
+	}
+
 protected:
 	///
 	/// \brief GetResult

src/Detector/tensorrt_yolo/YoloONNXv11_instance.hpp

@@ -7,6 +7,14 @@
 ///
 class YOLOv11_instance_onnx : public YoloONNX
 {
+public:
+	YOLOv11_instance_onnx(std::vector<std::string>& inputTensorNames, std::vector<std::string>& outputTensorNames)
+	{
+		inputTensorNames.push_back("images");
+		outputTensorNames.push_back("output0");
+		outputTensorNames.push_back("output1");
+	}
+
 protected:
 	///
 	/// \brief GetResult

src/Detector/tensorrt_yolo/YoloONNXv11_obb.hpp

@@ -7,6 +7,13 @@
 ///
 class YOLOv11_obb_onnx : public YoloONNX
 {
+public:
+	YOLOv11_obb_onnx(std::vector<std::string>& inputTensorNames, std::vector<std::string>& outputTensorNames)
+	{
+		inputTensorNames.push_back("images");
+		outputTensorNames.push_back("output0");
+	}
+
 protected:
 	///
 	/// \brief GetResult

src/Detector/tensorrt_yolo/YoloONNXv12_bb.hpp

@@ -0,0 +1,16 @@
+#pragma once
+
+#include "YoloONNX.hpp"
+
+///
+/// \brief The YOLOv12_bb_onnx class
+///
+class YOLOv12_bb_onnx : public YOLOv11_bb_onnx
+{
+public:
+	YOLOv12_bb_onnx(std::vector<std::string>& inputTensorNames, std::vector<std::string>& outputTensorNames)
+		: YOLOv11_bb_onnx(inputTensorNames, outputTensorNames)
+	{
+	}
+
+};