Skip to content
Branch: master
Find file History
Latest commit 5cf38da Aug 11, 2019
Permalink
Type Name Latest commit message Commit time
..
Failed to load latest commit information.
.gradle
.idea
app TensorFlowLite Aug 10, 2019
gradle/wrapper TensorFlowLite Aug 10, 2019
AndroidImageClassifier.apk
README.md Update README.md Aug 11, 2019
TensorFlow-Lite-Image-Classifier.iml TensorFlowLite Aug 10, 2019
TensorFlowLite.gif TensorFlowLite Aug 10, 2019
build.gradle TensorFlowLite Aug 10, 2019
gradle.properties TensorFlowLite Aug 10, 2019
gradlew
gradlew.bat
local.properties
output.json
settings.gradle TensorFlowLite Aug 10, 2019

README.md

TensorFlow-Lite-Image-Classifier

Creating an Android App using the pretrained Inception v3 Image Classifier

In this project, I imported the pretrained Inception v3 Image Classifier in an Android Studio App using TensorFlow Lite.

In order to create this project I took reference to the following sources:

Using this App, we first select the Inception Quantized Classifier (a compressed and less power hungry version of the full Inception Classifier). Successively, we then take a picture of the object we want to classify and finally we crop the picture (the Inception Classifier takes just squared images as inputs). Once we moved to the next window in the App, we can then click on the Classify Image button and our model will return the 3 most likely predictions.

If you are interested in testing out yourself the App, this is available to download here.

Some useful downloads used to create this project are:

You can’t perform that action at this time.