|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/computer-vision-bootcamptm-python-and-opencv/
课程评论：没有评论
课程名称：计算机视觉实训营与Python（OpenCV）- YOLO，SSD 课程概述：本课程主要围绕图像处理的基本概念展开，重点研究人脸检测和目标检测。这些主题在当今大热，因为这些学习算法可以应用于软件工程、犯罪调查等多个领域。自驾车（例如车道检测方法）在很大程度上依赖计算机视觉。随着深度学习和图形处理单元（GPU）在过去十年的发展，现在在实时视频中运行这些算法成为可能。 课程内容： 1. 图像处理基础：学习计算机视觉理论、像素强度值、卷积与滤波器、模糊核、锐化核及边缘检测。 2. 自驾车与车道检测：如何使用计算机视觉方法进行车道检测、介绍Canny算法和Hough变换。 3. Viola-Jones人脸检测算法：学习Viola-Jones方法、滑动窗口方法及如何在图像和视频中检测人脸。 4. 边缘梯度方向直方图（HOG）算法：介绍如何超越Viola-Jones算法，检测图像中的梯度与边缘、构建梯度方向直方图，并运用支持向量机（SVM）进行机器学习。 5. 基于卷积神经网络（CNN）的 approaches：讨论滑动窗口法的问题、区域建议和选择性搜索算法、区域卷积神经网络（C-RNNs）及其快速版本。 6. 你只需要看一次（YOLO v11）目标检测算法：学习YOLO方法、构建边界框、如何在图像中一次性检测对象、交并比（IOU）算法及非极大值抑制（NMS）实施YOLO11于图像与视频，并训练YOLO以适应自定义数据集。 7. 单次多框检测器（SSD）目标检测算法：了解SSD算法的主要思想、构建锚框、VGG16和MobileNet架构，并在实时视频中实施SSD。 8. 目标跟踪算法：DeepSORT、ByteTrack、BoTSORT算法及车辆计数算法的实现。 课程将系统讲解人脸识别和目标检测算法的理论背景，并逐步实现相关问题的解决。谢谢您参与课程，让我们开始吧！
This course is about the fundamental concept of image processing, focusing on face detection and object detection. These topics are getting very hot nowadays because these learning algorithms can be used in several fields from software engineering to crime investigation. Self-driving cars (for example lane detection approaches) relies heavily on computer vision. With the advent of deep learning and graphical processing units (GPUs) in the past decade it's become possible to run these algorithms even in real-time videos. So what are you going to learn in this course?Section 1 - Image Processing Fundamentals:computer vision theorywhat are pixel intensity valuesconvolution and kernels (filters)blur kernelsharpen kerneledge detection in computer vision (edge detection kernel)Section 2 - Serf-Driving Cars and Lane Detectionhow to use computer vision approaches in lane detectionCanny's algorithmhow to use Hough transform to find lines based on pixel intensitiesSection 3 - Face Detection with Viola-Jones Algorithm:Viola-Jones approach in computer visionwhat is sliding-windows approachdetecting faces in images and in videosSection 4 - Histogram of Oriented Gradients (HOG) Algorithmhow to outperform Viola-Jones algorithm with better approacheshow to detects gradients and edges in an imageconstructing histograms of oriented gradientsusing support vector machines (SVMs) as underlying machine learning algorithmsSection 5 - Convolution Neural Networks (CNNs) Based Approacheswhat is the problem with sliding-windows approachregion proposals and selective search algorithmsregion based convolutional neural networks (C-RNNs)fast C-RNNsfaster C-RNNsSection 6 - You Only Look Once (YOLO v11) Object Detection Algorithmwhat is the YOLO approach?constructing bounding boxeshow to detect objects in an image with a single look?intersection of union (IOU) algorithmhow to keep the most relevant bounding box with non-max suppression?implementation of YOLO11 with images and videostraining YOLO with custom datasetSection 7 - Single Shot MultiBox Detector (SSD) Object Detection Algorithm SDDwhat is the main idea behind SSD algorithmconstructing anchor boxesVGG16 and MobileNet architecturesimplementing SSD with real-time videosSection 8 - Object Tracking AlgorithmsDeepSORT object detection algorithmByteTrack algorithmBoTSORT algorithmimplementation of object trackingvehicle counting algorithmWe will talk about the theoretical background of face recognition algorithms and object detection in the main then we are going to implement these problems on a step-by-step basis.Thanks for joining the course, let's get started!
