Handwritten digit recognition using image processing
Download as PPTX, PDF8 likes7,377 views
The document presents a case study on handwritten digit recognition using image processing and neural networks. It discusses collecting handwritten digit images, preprocessing the images by cutting, resizing and extracting features, and then training a neural network using backpropagation to recognize the digits. The system aims to recognize handwritten digits for applications like signature, currency and number plate recognition. It concludes that understanding neural networks makes it easier to apply such intelligent recognition to machines.
Handwritten digit recognition using image processing
- 1. Handwritten Digit Recognition using Image Processing Team members Anita Maharjan(102/067/BEX) Chetana Moktan(108/067/BEX) (A presentation of a case study on title) 4/8/2014 1
- 2. Overview • Introduction • Objective • Implementation • System Overview • Neural Network • Conclusion 4/8/2014 2
- 3. Introduction • This working prototype system can detect handwritten digits from a scanned image of an input form by using Neural network technique. • very fast and effective as compared to old fashioned image pixel comparison methodology. 4/8/2014 3
- 4. Objective • To recognize handwritten digits in real works for autonomous machine processing. • To be familiar to Neural Networks 4/8/2014 4
- 5. Application Signature Recognition Currency Recognition Number Plate Recognition House no. Recognition4/8/2014 5
- 6. System Overview Handwritten digit image collection Image cut and store Image Slice Convert to standard size 4/8/2014 6
- 7. Calculate HH,VH,LDH,RDH Calculate GH=HH+VH+LDH+RDH Maintain Database Training Neural Network Digit Recognition 4/8/2014 7
- 8. Neural Network • Back Propagation Method • Hidden layer • Sigmoid activation function, O(i)=1/(1+e^(input)) 4/8/2014 8
- 10. Conclusion • Thus, understanding of neural networks • we have more control over its applications • now easy to implement such intelligence to identify objects into machines and computers • In order to cater our needs in the industrial applications 4/8/2014 10
- 11. Reference • Faisal Tehseen Shah, Kamran Yousaf*, “Handwritten Digit Recognition Using Image Processing and Neural Networks” • Youssef Es Saady, Ali Rachidi, Mostafa El Yassa, Driss Mammass , “Amazigh Handwritten Character Recognition based on Horizontal and Vertical Centerline of Character” , International Journal of Advanced Science and Technology, Vol. 33 4/8/2014 11