Five Unbelievable Open-Source Object Detection Projects Ready to Use in 2022

Object detection is a computer technology related to computer vision and image processing that deals with detecting instances of semantic objects of a certain class in digital images and videos.

In today’s article, we are going to talk about five of the open-source Object Detection projects to enhance your skills in the field of computer vision and image processing.

Note: In this article we are going to talk about some of the not-so-famous but really good open-source projects which you can use in your projects. To read more about each of them I recommend following the link given along the project.

Learning isn’t just about being more competent at your job, and it is so much more than that. Datacamp allows me to learn without limits.Datacamp provides you with the flexibility you need to take courses on your own time and learn the fundamental skills you need to transition to your successful career.Datacamp has taught me to pick up new ideas quickly and apply them to real-world problems. While I was in my learning phase, Datacamp *got me hooked with everything going on in the courses, from courses content and TA feedback to meetups events and the professor’s Twitter feeds.*Here are some of my favourite courses I would highly recommend you to learn from, whenever it fits your schedule and mood. You can directly apply the concepts and skills learned from these courses to an exciting new project at work or your university.

1. Data-scientist-with-python
2. Data-scientist-with-r
3. Machine-learning-scientist-with-r
4. Machine-learning-scientist-with-python
5. Machine-learning-for-everyone
6. Data-science-for-everyone
7. Data-engineer-with-python
8. Data-analyst-with-python
9. Big-data-fundamentals-via-pyspark

Coming back to the topic -

1.Imageai

Imageai** is a DeepQuestAI project. It is a free, open-source Python library used to build applications and systems with self-contained Deep Learning and Computer Vision capabilities with state-of-the-art Machine Learning Algorithms. It is made using *Python, OpenCV, Keras, and TensorFlow* frameworks.

It uses RetinaNet, YOLOv3,*,* and TinyYOLOv3** trained on the COCO dataset for *object detection, video object detection,, and object tracking*. It also supports image predictions using four different Machine Learning algorithms trained on the ImageNet-1000 dataset.

The best part of Imageai is that it allows you to train custom models for object detection and object recognition of your objects using your custom object dataset.

Currently, Imageai is Developed and Maintained by Moses Olafenwa and John Olafenwa, brothers

Image Source: https://github.com/OlafenwaMoses/ImageAI

2.AVOD

Aggregate View Object Detection is a project designed for 3D Object Detection for autonomous self-driving cars built on top of Python, OpenCV, and TensorFlow.

The authors trained the dataset for 3D object detection on Kitti Object Detection Dataset, and we compared the results with various other published methods on the *Kitti 3D object* and *BCV Benchmarks*. The Kitti dataset contains images of 8 different classes: Car, Van, Truck, Pedestrian, Person_sitting, Cyclist, Tram, Misc, and DontCare.

Source: https://github.com/kujason/avod

3.Nudenet

Nudenet is a free and open-source Neural Nets project used to detect and classify nudity in an image or video stream and selective censoring.

This project is built in Python and Keras. Self-hostable API service and a Python module are available for the direct implementation of the project. The latest version of Nudenet is trained on 160,000 *auto-labelled images with an accuracy of 93%*.

Installing the Nudenet library pip install nudenet, you can upload a photo/video and classify the image as:

• Safe — Image/video is not sexually explicit
• Unsafe — Image/video is sexually explicit

Source: An uncensored version of the following image can be found at https://i.imgur.com/rga6845.jpg (NSFW)

4. AI Basketball Analysis

AI Basketball Analysis** is an Artificial intelligence-powered**,** web app and API that you can use to analyze basketball shots and shooting poses built on top of the concept of object detection.

This project has mainly three features — shot analysis, shot detection, and detection API.

This project is implemented in Python using an open-source library, OpenPose (used to calculate the angle of elbow and knee). This project is built using the concept of transfer learning, and the based model used for training is Faster-RCNN which is pre-trained on the COCO dataset weights.

Source: https://ai-basketball-analysis.herokuapp.com/

5. Vehicle Counting

Vehicle Counting is an open-source project that focuses on Vehicle Detection, Tracking, and Counting. This project also provides predictions for the vehicle''s speed, colour, size, and direction in real-time using TensorFlow Object Detection API.

This project is implemented using Tensorflow, OpenCV, and Python, and the model used for vehicle detection is SSD with Mobilenet. Currently, this project is capable of classifying* five vehicles: Bus, Car, Cycle, Truck, and Motorcycle.*

Source: https://github.com/ahmetozlu/vehicle_counting_tensorflow

References

1. Imageai

@misc {ImageAI,
    author = "Moses and John Olafenwa",
    title  = "ImageAI, an open source python library built to   empower developers to build applications and systems  with self-contained Computer Vision capabilities",
    url    = "https://github.com/OlafenwaMoses/ImageAI",
    month  = "mar",
    year   = "2018--"
}

2. AVOD

• https://arxiv.org/abs/1712.02294
• https://github.com/kujason/avod

3. Nudenet

• https://github.com/notAI-tech/NudeNet

4. AI Basketball Analysis

• https://github.com/chonyy/AI-basketball-analysis

5. Vehicle Counting

@ONLINE{vdtct,
    author = "Ahmet Özlü",
    title  = "Vehicle Detection, Tracking and Counting by TensorFlow",
    year   = "2018",
    url    = "https://github.com/ahmetozlu/vehicle_counting_tensorflow"
}

If you enjoyed reading this article, I am sure that we share similar interests and are/will be in similar industries. So let’s connect via LinkedIn and Github. Please do not hesitate to send a contact request!