You are on page 1of 18

Machine Learning Internship

Fruit Detection Using


Convolutional Neural
Networks
Bontha Sarath Chandra (Y20CS027)

RVR&JC College Of Engineering

II/IV B.Tech. Computer Science and


Engineering

1
Abstraction
•Fruit detection is an important problem in computer vision and
has many practical applications, such as quality control in the
food industry and automated harvesting systems.
•This involves training a CNN model on a labeled dataset of fruit
images and using an optimization algorithm to adjust the
model's parameters.
•Our results show that the CNN was able to accurately classify the fruits
with an average accuracy of over 90%.

2
Introduction
•Convolutional Neural Networks (CNNs) are a type of Deep Learning
architecture that is particularly well-suited for image classification
tasks.
•In the context of fruit detection, the goal is to train a CNN model to
identify different types of fruits based on their images .
•Fruit Detection System can be established that can recognizes
various types of fruits from images that are captured using any
digital camera or smart phone from various places.

3
Fruits-360 Dataset
Training set size: 67692 images (one fruit or vegetable per
image).

Test set size: 22688 images (one fruit or vegetable per image).

Image size: 100x100 pixels.

4
Significance of this project
•The significance of this project lies in its potential applications and
the insight it provides into the capabilities of Deep Learning for
image classification tasks. Some of the key benefits and applications
of this project include:
Automated fruit sorting
Quality Control
Scientific research

5
Libraries / Technologies

Numpy
Pandas
TensorFlow
Keras
Matplotlib

6
Project Description
•The main objective of this project is to develop an automated fruit detection
system that can identify and classify different types of fruits based on their
images.
•This system should be able to accurately identify fruits, even when they are
partially obscured, and be able to handle variations in size, shape, and color. The
system should also be designed to be scalable and easily integrated into existing
food processing systems.
•The problem addressed by a fruit detection project using CNN is the need for
accurate and efficient identification of different types of fruits in a variety of
applications such as food processing , quality control .

7
Model Architecture

8
Convolutional Neural Networks Architecture

9
Feature Extraction

In Feature extraction part , the network will perform a series of convolutions


and pooling operations during which the Features are detected.

Classification

In case of Classification , the fully connected layers will serve as a classifier on


top of these Extracted Features.

10
11
12
13
14
15
16
Conclusion
In conclusion, a fruit detection project using Convolutional Neural
Networks (CNNs) is a significant project that has potential
applications in areas such as automated fruit sorting, quality control,
and scientific research.
The overall performance of model to detect fruit images is really
very well which has met a desired result of accuracy around 94%.
Overall, this project provides a good opportunity to explore
the capabilities of Deep Learning for image classification
tasks and to gain a deeper understanding of CNNs.

17
18

You might also like