search

KhEC / VIII-SEM-Project

Eight Semester Projects
1 stars 0 forks source link

Study 2-2 Research Papers #1

Open ErSKS opened 6 years ago

ErSKS commented 6 years ago

Each project member should study 2 relevant research papers and post the link of that papers with an exhaustive summary.

Project Member

jagdish4249 commented 6 years ago

Jagdish Duwal 1.Strategy to Extract Reliable Minutia Points for Fingerprint Recognition (link:https://drive.google.com/file/d/1TfiH8XIrkV0WFrnWfcHKTPCUNgCq5TGU/view)

2.Fingerprint Classification Using Convolutional Neural Networks and Ridge Orientation Images (link:https://zapdf.com/fingerprint-classification-using-convolutional-neural-networ.html)

PujanChangubhari commented 6 years ago

1.Neural Network based minutiae extraction for fingerprint verification system https://l.facebook.com/l.php?u=https%3A%2F%2Fdrive.google.com%2Ffile%2Fd%2F11uyslc3XjvfBwA-bvG1XGJuGAv5C99x2%2Fview&h=AT0OIodb0WNPgJK_2WCcq2BHbC1chsh9aRroqyysl57rQ35SQkgg773hVYgFgRz5pJONB_5F3wTQasLi39gYp8va-YbhKt75WSPOHRj0nzDxBSNoUBkQKvzHx1EBJGj_Oz35

The result of matching Accuracy is 91.6%.Extracting minutia features from fingerprint image using Feed-Forward Neural Network. It extract bifurcation minutiae points and core points.This method consist of five steps.They are:Image enhancement,thinning,minutiae extraction and graph representation.There are 3 layers.First input layer consist of 9 neurons,2nd input layer has 5 hidden layer and output layer as 1 neuron in which 1 for bifurcation and 0 for non bifurcation.Euclidean distance is used to match two fingerprints.

2.Fingerprint enhancement algorithm and performance evaluation https://l.facebook.com/l.php?u=https%3A%2F%2Fdrive.google.com%2Ffile%2Fd%2F1KF-QgE-FGSkSQmoVflqT9CAoDZVsyZ9e%2Fview&h=AT0OIodb0WNPgJK_2WCcq2BHbC1chsh9aRroqyysl57rQ35SQkgg773hVYgFgRz5pJONB_5F3wTQasLi39gYp8va-YbhKt75WSPOHRj0nzDxBSNoUBkQKvzHx1EBJGj_Oz35

Enhancement algorithm provides better ridge structures heavily depends upon the quality of input fingerprint image. Algorithm:Normalization :Local orientation estimation :Local frequency estimation :Region Mask estimation :Filtering Evaluation using good index Evaluation using verification process Both results better on applying fingerprint enhancement algorithm.

Kozu49 commented 6 years ago

1) Fingerprint Image Enhancement Algorithm and Performance Evaluation: (link: https://drive.google.com/file/d/1AyEXEVvuXXTgibk9mlzw07afQYKf7N1l/view)

->Fingerprint images get degraded and corrupted due to variations in skin and impression conditions Thus,image enhancement techniques are employed prior to minutiae extraction. ->The uniqueness of a fingerprint is exclusively determined by the local ridge characteristics and their rlationships.A total of 150 different local ridge characteristics have been identified. ->The set of minutiae types are restricted into two types,ridge ending and bifurcation. ->A good quality fingerprint typically contains about 40-100 minutiae but quality of fingerprint images is reduced due to noise,skin conditions etc. ->In a given digital fingerprint image,the region of interest can be divided into 3 categories: 1)Well-defined region(where ridges and valleys are clearly differentiated from one another) 2)Recoverable corrupted region(where ridges and valleys are corrupted by a small amount of creases,smudges) 3)Unrecoverable corrupted region(where ridges and valleys are corrupted by a sever amount of noise and distortion) ->Fingerprint enhancement can be conducted on either 1)Binary images(Better in comparison to gray level images) 2)Gray level images ->Fingerprint enhancement algorithm includes: 1)Normalisation 2)Segmentation 3)Orientation estimation 4)Ridge frequency estimation 5)Gabor filtering 6)Binarization and Thinning ->Performance Evaluation: i)Evaluation using Goodness Index ii)Evaluation using Verification Performance

2) Fingerprint Matching using Gabor Filters: (link: https://drive.google.com/file/d/1JGzhJ_PkUUUWKiuRh4GWUBJ4A1aunRqP/view)

->The proposed scheme uses a set of 16 Gabor filters,whose spatial frequencies correspppond to the average inter-ridge spacing in fingerprints,is usedto capture the ridge strength at equally spaced orientations. ->A circular tessellation of filtered image is then used to construct the ridge feature map. ->The ridge feature map contains both global and local details in fingerprint as a compact fixed length feature vector. ->Fingerprint matching is based on Euclidean distance between two corresponding feature vectors. ->The genuine accept rate of the gabor filter based matcher is observed to be 10% to 15% higher than that of minutiae-based matcher at low false accept rates. ->Core point detection: 1)Core point detection using poincare index 2)Core point detection using slope ->We have two core point locations from above two techniques.Optimal core point is calculatedby taking average of x-cor values and taking maximum of two y-cor values as maximum y-cor is more precise location of core point. ->Gabor filters optimally capture both local orientation and frequency information from fingerprint image.