Offline Signature Made Easy

Signature offline

welcome to our project powerful title each of them signature verification first of all we will see the importance of signature verification signature is an important aspect for authentication of individual every sectors like Bank public and government private office uses signature to authenticate injure or employ the individual side effect of signature in that they can be expected for the purpose of faking our documents authenticity so it is important to have some robust system to check genuine and forged signature so the basic steps involved in this project is firstly processing of images second is extracting feature or is creating issue I see that is stro rom of oriented gradient and fourth feature is training mode all interesting who know very walks for the court three views equal a cure oh they're going to use these levels on this project there's a shame to waste zero num-num poly antenna turning the section of put your own needs log in to try to our flesh so that we can move the trend agile of responds concealer to destroy the face Kazu you have all the content of type connealy access everything no the section of pole will have people sitting the persisting our input is awesome we take input our name comes near learn all this he positions he prepares awesome political system is a Jewish presence and facility noise signals Jerry machine trust and luminizing transcendental equation and then we are creating a kind of pause protagonist and a tuna for our speech today unwanted is this is from his cows and their mother this order food is with Kasich behind our unit so here in cellular network and inside this neural network they're really sigmoid function or edge our existence so as you can see once the ecosystem image and customization for plane is done they're calling this function here as the sigmoid function and passing push regulations it will run through intense and we're taking forever 21 this is a mystic emerald our train and inside this model we are passing this training data and this testing data you so US must retake the input we are creating a batches of engine so that process faster and by impact first the basis is created with using we assimilate the sigmoid function and we are trying to get the output and we are actually trying to him the time to do the air using back propagation was the relation part is done and every after running it for intents we are finding the test region so now we are going to run all the section one balloon so this is the importing library sections of a million worth of library I already have run this part so just skipping this now this is the main function where I'm actually in putting all the images testing and training I have to run this auto so now you're calling the main from some set will take some time because it will all the images and then it will train the model and again then we'll test the models that will take some time it will depend upon also on the alpha value and reports so so many difference what we take some more time or maybe some place the time will depend on that so you can see here yeah so my result gave 80 35 percent so actually it's idle value symbol 12 but it is fine 15 because of the model again we take some time and uh oh so now you can see this time they're showing this era again once you run the code I can see the sound of arrow mania we are very near to approach this kind of disturbance it very easier to us and again the disturbance in our output also let's conclude this session is following Holmes you you so a system where you know who the worst you became each and every forgery some is ill fortune he can mislead our system we are trying to make it at size and possible and you are watching this video