This is the most detailed course on Deep Learning using OpenCV’s DNN module out there, yes a complete 3-hour course that takes you from no background in DNN module, and teaches you all about it, including how to train custom classifiers and custom object detectors in TensorFlow and then deploy it in OpenCV module. You’ll also learn all the nitty-gritty details about this module in the course.
Download Code: https://app.monstercampaigns.com/c/zo...
Join for the Computer Vision Course:
https://bleedaiacademy.com/computer-v...
Blogpost for Each Section:
Section 1: https://bleedaiacademy.com/deep-learn...
Section 2: https://bleedaiacademy.com/training-a...
Section 3: https://bleedaiacademy.com/training-a...
If you want us to create AI and computer vision solutions for you,
then reach out to us here: https://bleedai.com/services/
--------------------------------------------------------------------------------------------------------------------------------------------------------------
00:00 Intro
01:08 Introduction to OpenCV's Deep Neural Network Module
01:45 What are the Benefits
03:39 A Bummer that You Must Know About
04:01 So Which Frameworks can be Used to Train Neural Networks?
05:22 Section 1 - Image Classification with DenseNet121
06:04 Import the Libraries
06:15 Load the Class Labels
08:05 Extract the Label
08:36 Initialize the DNN Module
11:48 Read an Image
12:19 Pre-process an Image
18:53 Input the Blob Image to the Network
19:12 Forward Pass
21:48 Apply Softmax Function to get Probabilities
25:27 Create Functions
26:57 Initialization Function
28:05 Classification Function
29:31 Initialize the Classifier
29:34 Classify Images Using the Function
30:12 Perform Real-Time Image Classification
32:55 Important Details Regarding the DNN Module
37:49 Section 2 - Training Custom Image Classifier with Tensorflow, Converting to ONNX and using it in OpenCV DNN module
38:48 (Part 1) Training a Custom Image Classifier with Tensorflow
38:51 Import the Libraries
42:07 Download and Unzip the Dataset folder
43:52 Check the number of images in each class
45:19 Generate Images
01:01:10 Visualize Images
01:04:11 Create the model
01:11:11 compile the model
01:11:13 Model Summary
01:19:28 Train The Model
01:23:22 Visualize Training Results
01:24:15 Test on new Examples
01:28:33 Save your model to Disk
01:29:29 (Part2) Converting Our Classifier to ONNX format
01:33:16 (Part 3) Using the ONNX model in the OpenCV DNN module
01:38:26 Create Functions
01:38:26 Initialization Function
01:38:36 Classification Function
01:42:32 Section 3 - Training Object Detector with Tensorflow Object Detection API
01:52:02 (Part 1) Environmental Setup
01:52:32 (Part 2) Installation & TFOD API Setup
01:55:11 Import Required Libraries
01:55:29 Clone Tensorflow Object Detection Model Repository
01:57:25 Install Tensorflow Object Detection API & Compile Protos
02:01:16 You can Check your installation of TFOD API by Running model_builder_test.py
02:02:06 Download the Support Folder required to run this Notebook
02:03:05 (Part 3) Data Collection & Annotation
02:04:41 Step 1: Download Youtube Video
02:05:49 Step 2: Split Video Frames and store it
02:12:55 Step 3: Annotate Images with labelImg
02:19:26 Step 4: Create a label Map file
02:21:04 Step 5: Generate TFrecords
02:39:09 (Part 4) Downloading Model & Configure it
02:39:09 (Part 5) Training and Exporting Inference Graph
--------------------------------------------------------------------------------------------------------------------------------------------------------------
If you enjoyed this video then don't forget to subscribe and like this video.
🎇 Subscribe to this Channel: https://bit.ly/Bleedsub
🎇 Subscribe to Bleed AI Blog: https://bleedaiacademy.com/subscribe/
🎇Check out our Courses: https://bleedaiacademy.com/courses/
Follow Bleed AI for more 👣 :
◼ Website: https://bleedaiacademy.com/
◼ Facebook Page: / bleedaiacademyy
◼ Linkedin: / bleedaiacademy
◼ Instagram: / bleedaiacademy
◼ Twitter: / bleedaiacademy
◼ Patreon: / bleedai
#imageclassification #objectdetection #deeplearning #crashcourse #opencv #tensorflow