bangla digit recognition using nn
DESCRIPTION
Bangla Digit Recognition Using NNTRANSCRIPT
Project Presentation ofBangla Digit Recognition using NN
Project Members:
Mohammad Nazmul Haque
Roni Shikder
Susmita Baidya
Ehsan Uddin Ahmed
Submitted to:Prof. Syed Akhter Hossain, Ph.DDaffodil International University
Presentation Outline1. Image Preprocessing
Image Acquisition Convert to Gray Scale Convert to Binary Image Identify maximum connected component Crop Individual character & make raw data Resize character into 5x7 ,applying fuzzy
technique
2. Create Neural Network3. Training the Network4. Testing The Network5. Performance Evaluation
1. Image Preprocessing Image Acquisition
Convert to grayscale image
1. Image Preprocessing Contt. make the image binary image &check for the
maximum connected component
1. Image Preprocessing Contt..
Crop individual characters & make raw data
sub-images have been resize to 50 by 70
bw_7050=imresize(bw2,[70,50]);
finding the average value in each 10 by 10 blocks
Atemp=sum(bw_7050((cnt*10-9:cnt*10),(cnt2*10-9:cnt2*10)));
1. Image Preprocessing Contt..
image can be down to 5 by 7 matrices, with fuzzy value
for cnt=1:7
for cnt2=1:5
Atemp=sum(bw_7050((cnt*10-9:cnt*10),(cnt2*10-9:cnt2*10)));
lett((cnt-1)*5+cnt2)=sum(Atemp);
end
end Resize Digit by 5x7 by average value
2. Create the NN Classification by simple FF-BP-NN The number of features, which is 5 by 7 = 35
Performance Goal : 0.01 Momentum constant : 0.95 Epochs : 5000
3 . Training the Network
Training Output Training Performance
TRAINGDX, Epoch 0/5000, SSE 100.468/0.1, Gradient 51.6634/1e-006TRAINGDX, Epoch 20/5000, SSE 39.1175/0.1, Gradient 0.755705/1e-006TRAINGDX, Epoch 40/5000, SSE 39.7742/0.1, Gradient 0.247939/1e-006TRAINGDX, Epoch 60/5000, SSE 39.8531/0.1, Gradient 0.1739/1e-006TRAINGDX, Epoch 80/5000, SSE 39.8688/0.1, Gradient 0.159411/1e-006TRAINGDX, Epoch 100/5000, SSE 39.8692/0.1, Gradient 0.160278/1e-006TRAINGDX, Epoch 120/5000, SSE 39.8604/0.1, Gradient 0.171435/1e-006TRAINGDX, Epoch 140/5000, SSE 39.827/0.1, Gradient 0.211754/1e-006TRAINGDX, Epoch 160/5000, SSE 39.628/0.1, Gradient 0.438358/1e-006TRAINGDX, Epoch 180/5000, SSE 34.2116/0.1, Gradient 2.91705/1e-006TRAINGDX, Epoch 200/5000, SSE 11.2205/0.1, Gradient 2.92171/1e-006TRAINGDX, Epoch 219/5000, SSE 0.0846207/0.1, Gradient 0.160478/1e-006TRAINGDX, Performance goal met
0 20 40 60 80 100 120 140 160 180 20010
-2
10-1
100
101
102
103
219 Epochs
Tra
inin
g-B
lue
Goa
l-Bla
ck
Performance is 0.0846207, Goal is 0.1
4. Testing the NN Testing Input:
Testing Output: 1 3 10 4 10 7 6 8 9 5 Testing Result Concatenated Individual Digits:
Performance : Case 1
Performance : Case 2
Performance : Case 3
Performance : Case 4
Performance SummaryParticulars Result
No of Training Digit 160
No of Testing Digit 40
No of Miss-Classification 2
Miss-Classified Digits ১, ৩ as ৯ , ০Performance (Success Rate) 95%
Future Scope Create NN based Bangla Number recognition
System Car’s Bangla Number Plate Recognition
System Road Speed Limit recognition system for
Bangla Road Sign
Thank You Any Questions / Suggestions ?