# Timber Identification CNN

## Introduction
This is the demo for my Final Year Project, it is a timber identification model. The model takes in an input image, then guesses the wood species of the given image.

The model is trained on microscopic wood cross sectional images, from 41 different species.

## Main Features
### How to use it
Simply go to the "Classification" tab, upload an image and press the submit button

### Sample Images
If you do not have any images, you may used the sample images provided in the "Samples" tab.

Simply go the the "Samples" tab, choose a species from the dropdown, and select an image by clicking on the "Submit" button

### Classification Output
After classification has completed, the predictions results will be shown. The header will be the main predictions. There will also be a dropdown, containing more details of the species. Note some species are more well documented than others, so the length of the details will vary greatly.

Below that are the confidence scores of the prediction, arranged in a descending order

### Prediction History
In the "History" tab, the results of previous predictions will be displayed, with the image uploaded and the species predicted.

Note that the history is only for the current session, so refressing the page will remove any saved states.