jinyup100

GSoC '20 : Real-time Single Object Tracking using Deep Learning (SiamRPN++)

Overview

Mentors : Liubov Batanina @l-bat, Stefano Fabri @bhack, Ilya Elizarov @ieliz
Student : Jin Yeob Chung @jinyup100
Mentors' Project Proposal : https://summerofcode.withgoogle.com/projects/#4979746967912448
Link to Pull Request : opencv/opencv#17647
Link to video summarising the experience : https://www.youtube.com/watch?v=D9G1vHqJCrc

Introduction

Recent interest in computer vision has led to a great advance in the development of visual trackers. Specifically, various applications of Kernelized Correlation Function (KCF) and deep learning have led to numerous implementations of single object trackers using publicly available libraries. Lately, there has been an increased focus on the function of convolutional features in developing visual trackers. In this particular project, I look to focus on the implementations of visual trackers

jinyup100

import numpy as np
import os
import onnx
import torch
import torch.nn as nn

# Class for the Building Blocks required for ResNet
class Bottleneck(nn.Module):
    expansion = 4

jinyup100

import numpy as np
import os
import onnx
import torch
import torch.nn as nn
import torch.nn.functional as F
import cv2
from torch.autograd import Variable

# Class for Region Proposal Neural Network

jinyup100

import math
import numpy as np
import os
import onnx
import torch
import torch.nn as nn
import torch.nn.functional as F
import cv2
from torch.autograd import Variable

jinyup100

import numpy as np
import os
import onnx
import torch
import torch.nn as nn
import torch.nn.functional as F
import cv2

# Class for RPN
class RPN(nn.Module):

jinyup100

import cv2
import math
import numpy as np
import os
import onnx
import torch
from torch.autograd import Function
from torch.autograd import Variable
import torch.nn as nn
import torch.nn.functional as F

jinyup100

import numpy as np
import os
import onnx
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.autograd import Variable

# Class for the Building Blocks required for ResNet