Flowers

Created By Sagar Saxena

06/14/18

Computer Vision Period 7

Programs Included In This Repository

cifarDemo.py - Classifies the cifar dataset. Found in TensorFlow Chapter 4
flowers10.py - Classifies 10 flowers using the same strategy as the Cifar Demo
train.py - Trains a Convulutional Neural Network Using Keras
Models.py - Models that can be trained by train.py
test.py - Tests a Single Image against a Saved Model
databases/database.py - Creates Directories of Images Into Pickled Cifar-like Batches
databases/imDownloader.py - Grabs Links from Google Images Using a Chrome WebBot
databases/linkDownloader.py - Converts all image links to images
databases/mousePy.py - Gives Control of the Mouse To Python

Databases Included In This Repository

databases/cifar-10-batches-py - Cifar 10 Batches
databases/mnist - Database of MNIST Handwritten Digits
databases/flowers10 - Database of 10 Flowers
databases/flowers91 - Database of 91 Flowers

Running Programs

Virtual Environment
    Enable: source/bin/activate
    Disable: deactivate
CifarDemo
    python3 cifarDemo.py
    python3 cifarDemo.py [Number of Steps] [Batch Size] [Print Skips]
    ex: python3 cifarDemo.py 5000 500 500
Flowers10
    python3 flowers10.py
    python3 flowers10.py [Number of Steps] [Batch Size] [Print Skips]
    ex: python3 cifarDemo.py 5000 200 500
Train
    python3 train.py -d [path to data batch] -m [path to model] -p [path to plot image] -e [Number of Epochs] -c [Number of Classes] -b [Size of Batch] -t [Type of Model]
    ex: python3 train.py -d databases/flowers10/data.pkl -m models/flowers10.model -p plotFlowers10.png -e 100 -c 10 -b 200 -t LeNet
    ex: python3 train.py -d databases/flowers10/data.pkl -m models/flowers10_c.model -p plotFlowers10_c.png -e 100 -c 10 -b 200 -t Cifar
    ex: python3 train.py -d databases/flowers91/data_batch_1.pkl -m models/flowers91.model -p plotFlowers91.png -e 100 -c 91 -b 200 -t LeNet
Models
    No Main Method
Test
    python3 test.py -m [path to model] -i [path to image] -l [path to labels file]
    ex: python3 test.py -m models/flowers10.model -i databases/flowers10/data/Columbine/3.jpg -l databases/flowers10/labels.txt
Database
    No Main Method
    python3
        from database import *
        genLabelFiles([Directory of Folders], [Path to Outfile], [Number To Skip])
        save_batches([Path To Directory of Folders], [Path To Label File], [Number of Batches], [Number of Images In Each Batch], [Directory To Save Data Batches])

        ex:
        genLabelFiles('./flowers10/data', './flowers10/labels.txt', 0)
        save_batches('./flowers10/data', './flowes10/labels.txt', 1, 600, './flowers10')
ImDownloader
    python3 imDownloader.py
    0
    __Name of Flower__

    #although other methods exist in this program (i.e. more than just method '0'), those methods should not be used as they were depreciated with the creation of linkDownloader.py
LinkDownloader.py
    No Main Method
    python3
        downloadFolder([Directory of Image Link Text Files], [Directory To Store Image Directories])
MousePy
    No Main Method
    Depreciated With imDownloader.py