You are on page 1of 25

IDENTIFICATION OF BIRD SPECIES USING

CONVOLUTION NEURAL NETWORK

Department Of Computer Science And Engineering


Under The Guidance Of
Mrs. S.Raga Deepthi, M.Tech
Assistant Professor Of CSE Department
V.CHANDRAKALA(18481A05L7)
S.PUJITHA (18481A05J4)
T.NAGA SAI LAKSHMI(18481A05L2)
V.LOKESWAR (18481A05M8)
ABSTRACT
Today, many species of birds are rarely found, and it is difficult to
classify bird species when found. For example, for different scenarios, birds come
with different sizes, forms, colors and from a human viewpoint with different angles
there are nearly 10 thousand living species of birds in the world with multifarious
characteristics and appearances. Bird watching is often considered to be an
interesting hobby by human beings in the natural environment. The human
knowledge over the species isn’t enough to identify a species of bird accurately, as it
requires lot of expertise in the field of Ornithology.
CNN is commonly applied for analyzing the visual imagery and image
Classification usually refers to taking an input of an image and classifying it to some
particular class, here in we have presented a CNN based classification model which
classifies the bird species given a bird image as input. The convolution neural
network model is capable of extracting the variant features based on size, shape and
color from the images and are hence capable of successful classification.
Introduction
 Numerous people visit bird sanctuaries to see various birds and to enjoy the
beautiful variants of colors and characteristics of the birds.
 People barely have the knowledge about the various species and hence cannot
easily distinguish the characteristics and the species name without the
expertise in the field of ornithology.
 we have presented a CNN based classification model which classifies the bird
species given a bird image as input.
 The convolution neural network model is capable of extracting the variant
features based on size, shape and color from the images and are hence capable
of successful classification.
 So we can use deep learning, It will help to identify the bird names and can
find the accuracy level.
Existing System

 BIRD behavior and population trends have become an important issue now a
days.
 Gathering and collecting information about birds requires huge human effort
as well as becomes a very costlier method.
 By using machine Learning we cannot train large amount of datasets.
 We cannot handle the images by using machine learning and it is also not
possible to finding its accuracy.
Proposed System
 In Proposed System we use deep learning for image processing to identify
bird species from digital images uploaded by an end-user.
 The proposed system could detect and differentiate uploaded images as birds.
 The overall accuracy is high for the training dataset using CNN model.
 This study ultimately aimed to design an automatic system for differentiating
among bird images with shared fundamental characteristics but minor
variations in appearance.
Objectives:
 

Bird species identification means predicting the bird species belongs to which category by using an
image. The identification can be done through image, audio or video.

 An audio processing technique makes it possible to identify by capturing the audio signal of
birds.

● But, due to the mixed sounds in environment such as insects, objects from real world, etc.
processing of such information becomes more complicated. Usually, human beings find images
more effective than audios or videos.
System Architecture:
CNN consists of four layers:

 1.Feature extraction part :


 2. Classification part
 3.Pooling
 4.Fully Connected Layer
1.Feature extraction part : features are detected when the network performs a series of the
convolutional and pooling operation.
2. Classification part : extracted features are given to a fully connected layer which acts as a
classifier CNN consists of four layers: convolutional layer, activation layer, pooling layer, and
fully connected. The convolutional layer allows extracting visual features from an image in
small amounts.

3.Pooling : Pooling is used to reduce the number of neurons from the previous convolutional
layer but maintaining the important information. The activation layer passes a value through a
function that compresses values into range.

4. Fully connected layer : connects a neuron form one layer to every neuron in another layer.
As CNN classifies each neuron in-depth, so it provides more accuracy [5-6].
Methodology:
Main purpose behind this experiment is to identify the image of bird and classify in to a concern
species by considering following objectives.
=>This study developed a platform that uses deep learning for image processing to identify bird
species from digital images uploaded by an end-user. The proposed system could detect and
differentiate uploaded images as birds. With an overall accuracy is high for the training dataset
using CNN model. =>This study ultimately aimed to design an automatic system for
differentiating among bird images with shared fundamental characteristics but minor variations in
appearance.
A.Browse/upload Image :
Caltech-UCSD.Caltech-UCSD Birds 200(CUB- 2002011) is a well-known bird image dataset with 200
category images. The dataset includes birds found mostly in North America [6]. Caltech-UCSD Birds
200 consists of 11,788 pictures and annotations such as 312 binary attributes, 15 component positions,
1 bounding box.
B. Pre-processing:
Pre-processing developed a gray scale image dataset that is used to pixel-by-pixel image recognition
and image size reductions. [2] Then, these functions are aggregated and forwarded to the classifier.
This increased processing time while retaining quality of the image.
C. Deep learning :
A DCNN is composed of different layers of convolution. Different alignment or features such as head,
body, color, beak, form, and whole bird image are considered to give maximum classification accuracy
This alignment is constructed through a deep network to extract multi functions.
D.Classification: :

In this analysis, the GoogLeNet framework has been used to identify the images.
Tensorflow is a software library which is created by Google and which is an open source
in nature It allows developers to monitor each neuron (node) in order to adjust the
parameters to achieve the desired output [12]. Tensorflow has a number of built-in image
classification libraries. Tensorflow [3] produces an autograph consisting of sequence of
processing nodes and retraining the dataset to achieve greater recognition accuracy
E. Evaluation :

To produce possible results, the input will be compared with the trained dataset an
autograph, consisting of nodes that ultimately form a network, is created during
classification. A score sheet is generated on the basis of this network and will be
created with the aid of the score sheet output [4].
Flow Diagram:
Use Case Diagram:

upload image

Run DCNN Algorithm

User

View Identified Species

View Score card


Sequence Diagram:

user Database

upload image

Run DCNN Algorithm

View Identified species

View Score card


Activity Diagram:

Start

Choose Image

Upload Image

iRun DCNN Algorithm

View Identified Species

View Score card

Predicting with the model

Result
DATA SET:
Execution
selecting an image from the data.
Uploading image
RUN DCNN ALGORITHM & VIEW IDENTIFIED SPECIES:
To know the species name of uploaded bird, screen we got 5 related birds images
of uploaded image and we can see the species name of bird on title bar of image.
So by uploading any image we can know the name of bird. You can upload any
image and get it name and uploading image name should be as integer value.
View Score Graph:

we got matching score of all 5 related birds and in above graph x-axis represents name of bird and y-axis represents matching score.
CONCLUSION :
This model helps building applications that helps tourist who go onto bird
sanctuaries identify the bird species by just capturing a picture of a bird and
uploading it as input to the model. As many species of birds have become
endangered and are near to extinction many people have no knowledge about
the species which are few in number, Thus application built using this model
may be helpful in identifying the endangered species and help society in
spreading awareness about the need of all the species for balance in the
nature. As the model implies the knowledge of Deep Convolution neural
networks, we can infer that the CNN is the best algorithm for analyzing the
visual imagery and image Classification.

You might also like