In this tutorial, we will cover everything you need to know about Tensorflow Object Detection in 5 hours with Python. We will be diving deep into the concepts of object detection using Tensorflow, covering different models, training techniques, and projects to help you get hands-on experience.
Having a basic understanding of Python programming and machine learning concepts will be helpful, but not required as we will cover everything from scratch.
Here’s a breakdown of the tutorial:
1. Introduction to Object Detection:
We will start by understanding what object detection is and why it is important in the field of computer vision. We will then introduce Tensorflow, an open-source machine learning library developed by Google, and discuss why it is the preferred choice for object detection tasks.
2. Setting up the Environment:
We will guide you through setting up your Python environment with all the necessary libraries, including Tensorflow, OpenCV, and NumPy. We will also show you how to install the Tensorflow Object Detection API, which provides pre-trained models and tools for training custom models.
3. Understanding Object Detection Models:
We will explain the different types of object detection models, including Single Shot Multibox Detector (SSD), Region-based Convolutional Neural Network (R-CNN), and You Only Look Once (YOLO). We will discuss the pros and cons of each model and when to use them.
4. Training a Custom Object Detection Model:
We will walk you through the process of training a custom object detection model using the Tensorflow Object Detection API. You will learn how to prepare your dataset, annotate images, and configure the model for training. We will also discuss techniques for improving the performance of your model, such as data augmentation and transfer learning.
5. Evaluating and Testing the Model:
Once your model is trained, we will show you how to evaluate its performance using metrics such as Precision, Recall, and Mean Average Precision (mAP). We will also demonstrate how to test the model on new images and videos to detect objects in real-time.
6. Projects:
We will conclude the tutorial with three hands-on projects that will help you apply the concepts you have learned. The projects will include:
– Building a Real-time Object Detection System using a webcam
– Detecting Objects in Image and Video Files
– Customizing a Pre-trained Object Detection Model for a Specific Task
By the end of this tutorial, you will have a solid understanding of Tensorflow Object Detection and the skills to build your own object detection models for various applications. We hope you enjoy this full course and find it helpful in your journey towards becoming a proficient computer vision developer. Let’s get started!
Wilson Carol Johnson Amy Thompson Robert
Taylor Margaret Martinez Mark Lee Daniel
Garcia Lisa Thompson Betty Thompson Margaret
Rodriguez David Williams Donna White Maria
Moore Jeffrey Allen Linda Jackson Helen
Williams Deborah Clark Anthony Wilson Jeffrey
Young Gary White Dorothy Martin Linda
I am running this on WSL
When I am running the verification script it is giving this error – AttributeError: module 'keras._tf_keras.keras.layers' has no attribute 'experimental'
After searching the internet, I changed the version of TF to 2.13.1 and matplotlib to 3.5.0.
Then on running the verification script it geave an OK.
Now when I am trying to import object_detection, it is giving Error – ModuleNotFoundError: No module named 'object_detection'
At this point I don't know what to do
very good tutorial. clear with explanations and code and all step by step good job.
I don't know if you checked the comments but I'm a beginner in python and I can't solve this problem :😄
In the update config for transfer learning part, my code shows a list of error saying "If this call came from a _pb2.py file, your generated code is out of date and must be regenerated with protoc >= 3.19.0.
If you cannot immediately regenerate your protos, some other possible workarounds are:
1. Downgrade the protobuf package to 3.20.x or lower.
2. Set PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python (but this will use pure-Python parsing and will be much slower)."
It would be great if you could help me with this!
When running the verification script I get the error "AttributeError: module 'keras._tf_keras.keras.layers' has no attribute 'experimental'". Currently running on Python 3.11.4 and TF 2.17
Thanks 👍
READ BEFORE WATCHING THIS!!!
No disrespect to creator, but as of 2024 This tutorial is SEVERELY outdated. The module versions given in the scripts are too version specific and almost nothing integrates well with one another.
It was soo hard just to get this working. Trust me, I don't want anyone to suffer through that. There's no help in the internet, as well as his discord sever. It's dead. I would advise you to not waste your time on this
Going through the same thing but protoc and protobuf has taken the charge to not make this complete. Ive done so many research and so many things but somehow, this protoc and protobuf things are not working at all. Help me out here man. There is always an error of cannot import anchor_generator_pb2 from object_detection. Have had all the stackoverflow pages now for this thing but there is no solution. Just help me out here man!
Edit : For some fucking reason, all my files downloaded in my environment are somehow reflected inside my global file download in python main directory in C drive. And for some fucking reason all the correct files from my environment doesn't matter at all but there are somehow some missing files in my global directory so there is an error for that fucking unusual thing even though everything done was done in my environment. Help me out man I've somehow been stuck here at this fucking thing for more than a week.
i get some problems with the libraries, just have to install it in cmd
Thank you sooo much for this video, i wana ask what python version you used throughout this project
I have collected thousands of datasets of images…now labeling them manually is impossible how can that be achieved??? plz help me..
Hmm, pip does not support egg file and setup installer or something. I got error when installing Tensorflow object detection for Python at 1:47:53 which can not verify my installation
i am facing an error AttributeError Traceback (most recent call last) Cell In[95], line 1 —-> 1 config = config_util.get_configs_from_pipeline_file(files['PIPELINE_CONFIG']) File ~Tensorflow Object DetectionTFODCourseobject_detectionLibsite-packagesobject_detectionutilsconfig_util.py:96, in get_configs_from_pipeline_file(pipeline config_path, config override) 82 """Reads config from a file containing pipeline_pb2.TrainEvalPipelineConfig. 83 84 Args: (…) 93 corresponding config objects. 94 """ 95 pipeline_config = pipeline_pb2.TrainEvalPipelineConfig() —> 96 with tf.gfile.GFile(pipeline_config_path, "r") as f: 97 proto_str = f.read() 98 text_format.Merge(proto_str, pipeline_config) AttributeError: module 'tensorflow' has no attribute 'gfile'
Hi Nick, your videos are amazing!
I'm relatively new at this stuff and starting a while after you made it so software has probably changed since then. I'm getting a lot of Depreciated errors – C:Tensorflow Object DetectionTFODCoursetfodLibsite-packagessetuptools_distutilscmd.py:66: SetuptoolsDeprecationWarning: setup.py install is deprecated.
Please avoid running “setup.py“ directly.
Instead, use pypa/build, pypa/installer or other
standards-based tools.
I've looked for solutions but seems to be something out of my control?
Any advice on getting around this?
1:28:53 timestamp