Groups keyboard shortcuts have been updated
Dismiss
See shortcuts
TUTORIAL #4 - ACCURATE HANDWRITTEN DIGITS RECOGNITION WITH MNIST AND TENSORFLOW
530 views
Skip to first unread message
TensorFlow
Nov 21, 2015, 2:49:00 AM11/21/15
to TensorFlow End Users - GETTING STARTED, TUTORIALS & HOW-TO'S
TUTORIAL #4
*** TUTORIAL TITLE ***
TENSORFLOW, MNIST AND YOUR OWN HANDWRITTEN DIGITS
*** TUTORIAL DESCRIPTION ***
"There are a lot of articles about MNIST and how to learn handwritten digits. So this one will be just another one?
Nope, the author Ole Kröger will use the newest available library Tensorflow by Google.
Although TensorFlow.org already has a MNIST tutorial Ole Kröger didn't find any article which describes how to recognize the digits as we normally write them.
All digits inside MNIST look really cool, however no one writes with a perfect white pen on a black background.
You will use your camera and make a photo where the digit isn't in the center.
If you use that as the input for your neural net, you will get some random results.
The last part of every article about MNIST is about the accuracy which is something around >85% and you will get something like 10% (random).
HOW TO GET THIS ACCURACY WITH YOUR OWN HANDWRITTEN DIGITS?
*** TUTORIAL SYLLABUS ***
1 - The MNIST dataset - a small overview
2 - Tensorflow - Library for machine learning by Google
3 - Preprocessing
4 - CHECKING ACCURACY
*** TRY IT ***
http://openmachin.es/blog/tensorflow-mnist
https://github.com/openmachinesblog/tensorflow-mnist
*** TUTORIAL TITLE ***
TENSORFLOW, MNIST AND YOUR OWN HANDWRITTEN DIGITS
*** TUTORIAL DESCRIPTION ***
"There are a lot of articles about MNIST and how to learn handwritten digits. So this one will be just another one?
Nope, the author Ole Kröger will use the newest available library Tensorflow by Google.
Although TensorFlow.org already has a MNIST tutorial Ole Kröger didn't find any article which describes how to recognize the digits as we normally write them.
All digits inside MNIST look really cool, however no one writes with a perfect white pen on a black background.
You will use your camera and make a photo where the digit isn't in the center.
If you use that as the input for your neural net, you will get some random results.
The last part of every article about MNIST is about the accuracy which is something around >85% and you will get something like 10% (random).
HOW TO GET THIS ACCURACY WITH YOUR OWN HANDWRITTEN DIGITS?
*** TUTORIAL SYLLABUS ***
1 - The MNIST dataset - a small overview
2 - Tensorflow - Library for machine learning by Google
3 - Preprocessing
4 - CHECKING ACCURACY
*** TRY IT ***
http://openmachin.es/blog/tensorflow-mnist
https://github.com/openmachinesblog/tensorflow-mnist
Rukamesh Kumar
Jun 9, 2016, 4:29:34 AM6/9/16
to TensorFlow End Users - GETTING STARTED, TUTORIALS & HOW-TO'S
Hi,
I am new to tensorflow and i want to make a model for image recognition.it should be read the input from csv file in which we have image path with
their corresponding labels.
I am new to tensorflow and i want to make a model for image recognition.it should be read the input from csv file in which we have image path with
their corresponding labels.
Jay Vala
Feb 24, 2017, 8:20:38 AM2/24/17
to TensorFlow End Users - GETTING STARTED, TUTORIALS & HOW-TO'S
All the turorials shows what the accuracy was but no one has ever tried to see what were the feeds to the first layer during testing phase and what it predicted(labels), what I think is if we have some kind of data on what it does inside we would be able to fine tune it to a level that it is more accurate.
If there was a was to create a matrix where we would be able to see what were the feeds input to the networks and what it predicted, rather than only 10 or 15 results we will be able to get more results and not get restricted to a few datapoints.
Reply all
Reply to author
Forward
0 new messages