0% found this document useful (0 votes)
134 views

Object Detection Using Image Processing

This document discusses object detection using image processing and machine learning. It describes how object detection models can detect and classify objects from images and videos using techniques like YOLO and R-CNN. The goal is to build accurate machine learning models that can detect objects as well as existing models. Sample code uses Python, PyTorch and TensorFlow to implement real-time object detection with YOLO, which looks at the entire image only once to predict object boundaries and classes quickly. Training models with large datasets can improve accuracy for tasks like robotics, tracking and threat detection.

Uploaded by

Sumit Singh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
134 views

Object Detection Using Image Processing

This document discusses object detection using image processing and machine learning. It describes how object detection models can detect and classify objects from images and videos using techniques like YOLO and R-CNN. The goal is to build accurate machine learning models that can detect objects as well as existing models. Sample code uses Python, PyTorch and TensorFlow to implement real-time object detection with YOLO, which looks at the entire image only once to predict object boundaries and classes quickly. Training models with large datasets can improve accuracy for tasks like robotics, tracking and threat detection.

Uploaded by

Sumit Singh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 17

Object

Detection
Using Image
Processing
DESCRIPTION

  It's a machine learning model that detects and classifies objects from images, stored
videos or real-time videos.

  One main objective is to reach a good accuracy so that the model recognizes and
classifies an object correctly.

 Object detection is a very important task in image processing and computer vision. It
forms the basis of robotics, tracking systems and threat detection systems.
PROBLEMS

❖  To build a machine learning model which detects and classifies objects from images and videos.

❖  To compare the accuracies of this machine learning model and the pre-existing models, and draw
the necessary conclusions.
SOLUTION
  Humans can detect and categorize objects easily but the same thing can't be said for
computer systems.   

  By using machine learning and deep learning it will be much easier to describe the
objects and classify them using multiple techniques like Background Subtraction
Method, Shadow Detection Technique and Template Matching.

❖ By training the model with sufficient datasets we can increase the accuracy of the
model.
A SAMPLE EXAMPLE OF REAL-TIME
OBJECT DETECTION
Concept and Method used:
o Machine Learning
o Deep learning
o YOLO — You Only Look Once
o R-CNN — Region With Convolutional Neural Networks
o Faster R-CNN

Language and library:


o Python
o PyTorch
o TensorFlow
YOLO – YOU ONLY LOOK ONCE

 Many of object detection systems need to go through the image more than one time to be
able to detect all the objects in the image, or it has to go through two stages to detect the
objects. YOLO doesn’t need to go through these boring processes .It only need to look
once at the image to detect all the objects and that is why they chose the name (You Only
Look Once) and that is actually the reason why YOLO is a very fast model.

 YOLO VERSIONS –
 YOLO(FIRST VERSION)
 YOLO V2
 YOLOv3
YOLO WORKING
 we crop our original photo. YOLO divides the input image into an S×S grid. Each grid cell predicts
only one object.
 For each grid cell,
• it predicts B boundary boxes and each box has one box confidence score,
• it detects one object only regardless of the number of boxes B,
• it predicts C conditional class probabilities (one per class for the likeliness of the object class).
 YOLO uses 7×7 grids (S×S), 2 boundary boxes (B) and 20 classes (C).
 YOLO uses sum-squared error between the predictions and the ground truth to calculate loss. The loss
function composes of:
• the classification loss. (the classification loss at each cell is the squared error of the class conditional probabilities)
• the localization loss (errors between the predicted boundary box and the ground truth).
• the confidence loss (the objectness of the box).
Benefits of YOLO

 Fast. Good for real-time processing.


 Predictions (object locations and classes) are made from one single network. Can
be trained end-to-end to improve accuracy.
 YOLO is more generalized. It outperforms other methods when generalizing from
natural images to other domains like artwork.
 Region proposal methods limit the classifier to the specific region. YOLO
accesses to the whole image in predicting boundaries. With the additional context,
YOLO demonstrates fewer false positives in background areas.
 YOLO detects one object per grid cell. It enforces spatial diversity in making
predictions.
Architecture Diagram
The objects detected by YOLO
TEAM GUIDE
⮚Rajesh Kumar  S. Janarthan Sir
⮚Satyam Singh
⮚Rajat Kumar Singh
THANK YOU

You might also like