There are many questions online about how to get started in computer vision. Most authors emphasize that practical projects and theoretical learning are equally important, as over-focusing on theory while neglecting practice can easily lead to the path of ‘from beginner to giving up.’Although there is a consensus on this, there is no systematic summary online about which projects beginners in CV should start with. Therefore, this article will organize some suitable projects for newcomers in the field of computer vision, categorized into five main tasks: Object Detection, Object Tracking, Image Segmentation, Image Classification, and Image Generation. A brief introduction and links to these projects will also be compiled in the article. Since external links cannot be inserted, you can click on ‘Read the original text’ at the end of the article. I hope this article can truly help many beginners in the field of computer vision!
Object Detection
1. License Plate Recognition using YOLOv4, OpenCV, and Tesseract OCR
Project Link: github.com/theAIGuysCod
2. YOLO V5 Object Detection using Pytorch
YOLOv5 is a series of object detection architectures and models pre-trained on the COCO dataset, representing Ultralytics’ open research on future visual AI methods, which incorporates the experience and best practices gained from thousands of hours of research and development.Demo Link: youtube.com/watch?Project Link: github.com/ultralytics/Colab Link: colab.research.google.com
3. Face Mask Detection using Python, Keras, OpenCV, and MobileNet | Detect Masks in Real-Time Video Streams
3. Unsupervised Learning Video Segmentation Method for Effective Learning of Spatiotemporal Correspondences
Demo Link: Unsupervised Learning Video Segmentation Method! Effective Learning of Spatiotemporal Correspondences | NeurIPS 2021Project Link: GitHub – visinf/dense-ulearn-vos: Dense Unsupervised Learning for Video Segmentation (NeurIPS*2021)
4. Multi-Modal Transformer for Video Segmentation
Demo Link: CVPR2022 | Multi-Modal Transformer for Video Segmentation Effects StunningProject Link: github.com/mttr2021/MTT
5. Query-Based Instance Segmentation
Demo Link: Huazhong University of Science and Technology & Tencent Proposed: Query-Based Instance Segmentation! Code has been Open-Source | ICCV2021Project Link: github.com/hustvl/Query
6. UNet for Medical Image Segmentation
Demo Link: A Step-by-Step Guide to Using UNet for Medical Image SegmentationProject Link: github.com/肆十二/unet_42
Image Classification
1. Image Classification using Keras, TensorFlow | Cat and Dog Prediction | Convolutional Neural Network
3. Image Classification With Deep Learning And TensorFlow: Intro Project
In this project, we will complete an end-to-end deep learning project using TensorFlow and Keras. We will read a dataset of dog images and train a convolutional neural network to classify them by breed. Finally, you will learn how to use Keras to train and optimize neural networks, and you will also learn how to process images using Python.Demo Link: youtube.com/watch?Project Link: github.com/dataquestio/
4. Real Python Neural Network Tutorial (Image Classification w/ CNN) | TensorFlow & Keras
5. Neural Texture Extraction and Distribution for Controllable Character Image Synthesis
Demo Link: CVPR2022 | Neural Texture Extraction and Distribution for Controllable Character Image Synthesis! Open SourceProject Link: github.com/RenYurui/Neu
