xeye 1.0

Create datasets for computer vision applications

Xeye

• Xeye
  • Installation
  • Xeye datasets for deep learning
  • Xeye functionalities
  • Create a dataset with full terminal UI (Dataset)
    • Other useful methods
    • Script example
  • Create a dataset with fast UI (FastDataset)
  • Create a dataset with manual UI (ManualDataset)
    • How to take pictures manually
  • Build datasets from different npz files (BuildDataset)
    • Script example

Xeye is a package for data collection to build computer vision applications based on inferential results of deep learning models. The main reasons to use Xeye are:

• Create a dataset using only a laptop and its integrated camera (or alternatively an external USB camera);
• Create a dataset already structured like the mnist;
• Create a dataset that can be used for building models with Tensorflow or Pytorch.

---

Installation

To install the package,

pip install xeye

---

Xeye datasets for deep learning

In the xeye-notebooks repository, you can find examples of deep learning model implementations based on datasets produced by the Xeye package (made with Tensorflow or Pytorch frameworks).

• Binary dataset: containing two types of grayscale images (with labels: 0=keyboard, 1=mouse).
• MultiLabel dataset: containing three types of rgb images (three types of security cameras with labels: 0=dome, 1=bullet, 2=cube)

Additionally, the xeye-notebooks repository contains examples of scripts that use the Xeye package to build datasets (examples link).

---

Xeye functionalities

The Xeye package includes three major approaches (classes) for creating a dataset from scratch: Dataset, FastDataset, and ManualDataset.

• Dataset: Uses the full UI terminal interface.
• FastDataset: Uses the constructor with all the specifications of the dataset.
• ManualDataset: Same as FastDataset, but every image is shot manually one at a time.

Additionally, the package provides a method for combining datasets created with the BuildDataset class.

---

Create a dataset with full terminal UI (Dataset)

First of all, load the module datapipe from the package:

import xeye

then initialize the instance like this

data = xeye.Dataset()

set the parameters related to the images with the setup method

data.setup()

the execution of this method causes the starting of the user interface in the terminal

--- CAMERA SETTING ---
Select the index of the camera that you want to use for creating the dataset: 1

the setup method arises multiple questions that set the parameters' values

--- IMAGE SETTINGS ---
Num. of types of images to scan: 2
Name of image type (1): keyboard
Name of image type (2): mouse
Num. of frames to shoot for every image type: 10
Single frame HEIGHT: 720
Single frame WIDTH:  720
num. of waiting time (in sec.) between every frame: 0

Precisely the questions refer to:

• Select the index of the camera that you want to use for creating the dataset: Generally 0 for integrated camera, 1 for USB external camera.
• Num. of types of images to scan: Answer 2 if you want to create a dataset with two objects (e.g. keyboard and mouse). In general, answer with the number of object types to include in your dataset.
• Name of image type: Insert the name of every specific object you want to include in the dataset. The setup method creates a named folder for every image type.
• Num. of frames to shoot for every image type: Select the number of images you want to shoot and save them in every object folder.
• Single frame HEIGHT: Frame height values.
• Single frame WIDTH: Frame width values.
• Num. of waiting time (in sec.) between every frame: e.g 0.2 causes a waiting time of 0.2 seconds between every shoot.

After the parameters setting, you can invoke the method to start shooting images. Datapipe module provides two different formats of images:

• Grayscale image with the gray method;
• Color image with the rgb method.

Let's produce a dataset based on RGB images with the rgb method:

data.rgb()

In the terminal keypress [b], to make photos for the image types passed to the setup method.

--- START TAKING PHOTOS ---
Press [b] on the keyboard to start data collection of image type [keyboard]
b
Press [b] on the keyboard to start data collection of image type [mouse]
b

On the directory of the script, you can find the folders that contain the images produced by the rbg method (e.g. keyboard folder and mouse folder).

Images collected in the folders can be used for building datasets like the mnist. The first approach to achieve this result is calling the compress_train_test method:

data.compress_train_test()

That produces the following output in the terminal window

--- DATASET SETTING ---
percentage of images in the test set: 0.2

In which you can select the portion of images to use in the train and test datasets (write a value between (0,1)). By doing so, the method produces a .npz file formed by these specific tensors:

• Train set:
  • X_train: Matrices/tensors of every single image in the train dataset;
  • y_train: Classes (ordinal values) are associated with every image in the train dataset.
• Test set:
  • X_test: Matrices/tensors of every single image in the test dataset;
  • y_test: Classes (ordinal values) are associated with every image in the test dataset.

(matrices for grayscale images: [Height $\times$ Width $\times$ 1], tensors for RGB images:[Height $\times$ Width $\times$ 3]).

An alternative approach is represented by the use of the compress_all method

data.compress_all()

In this case, the images are united in a unique tensor that contains all the frames produced previously.

• Unique tensor:
  • X: Matricies/tensors of every single image produced;
  • y: Classes (ordinal values) are associated with every image produced.

Finally, you can use the just_compress method to create a unique tensor with all the images produced.

data.just_compress()

In the terminal, you have to insert the dataset’s name

--- DATASET SETTING ---
Select a name for the compressed file: batch_test

If you pass 0, by default the dataset will be save with the name dataset_raw.npz. The dataset produced by this method can be used by the class BuildDataset to put together more .npz files and create a dataset like the mnist.

Other useful methods

• preview: Open the camera stream to check the framing.
• var_control: Print the values of the parameters set with the setup method.

--- PARAMETERS CONTROL ---
camera index: 1
num. of images types: 2
labels of images types: ['keyboard', 'mouse']
num. of images for types: 20
Single frame HEIGHT: 720
Single frame WIDTH:  720
waiting time between frames: 0.0
percentage of images in train dataset: 0.2

Script example

Example of a script to use the Dataset class:

import xeye
data = xeye.dataset()
data.setup()
data.preview()
data.rgb()
data.compress_train_test()
data.compress_all()
data.just_compress()

---

Create a dataset with fast UI (FastDataset)

The FastDataset class provides a faster way to use the datapipe module. Unlike the Dataset class, it does not have a complete terminal UI to guide you through the dataset construction process. Instead, you simply pass the parameters to the class and call the necessary methods.

import xeye
# define parameters values
index = 0
img_types = 2
label = ['keyboard', 'mouse']
num = 20
height = 100
width = 100
standby_time = 0
# percentage of images in the test set 
perc = 0.2
data = xeye.FastDataset(index = index, img_types = img_types, label = label, num = num, height = height, width = width, stand_by_time = standby_time)
data.preview()
data.rgb()
data.compress_train_test(perc = perc)
data.compress_all()
data.just_compress("batch_test")

The parameters passed to the class FastDataset:

• index: Generally 0 for integrated camera, 1 for USB external camera.
• img_types: Numbers of object types that you want to include in your dataset.
• label: List of object names to include in the dataset. The setup method creates a named folder for every image type.
• num: Select the number of images you want to shoot and save them in every object folder.
• height: Frame height values.
• width: Frame width values.
• standby_time: e.g 0.2 cause a waiting time of 0.2 seconds between every shoot.

For split images in the train and test dataset, pass a value between (0,1) to the perc parameter of the compress_train_test method

• perc: the portion of images to use in the test dataset, write a value between (0,1).

If you don't pass any value to the just_compress method, the dataset will be saved with the name dataset_raw.npz.

---

Create a dataset with manual UI (ManualDataset)

The ManualDataset class is how you can build a dataset by taking pictures manually.

import xeye
# define parameters values
index = 0
img_types = 2
label = ['keyboard', 'mouse']
num = 20
height = 100
width = 100
standby_time = 0
# percentage of images in the test set 
perc = 0.2
data = xeye.ManualDataset(index = index, img_types = img_types, label = label, num = num, height = height, width = width)
data.preview()
data.rgb()
data.compress_train_test(perc = perc)
data.compress_all()
data.just_compress("batch_test")

As you can see in the code snippet, the ManualDataset class works like the FastDataset class. The only difference is the absence of the standby_time option, which is no more necessary in this case.

The parameters passed to the class ManualDataset:

• index: Generally 0 for integrated camera, 1 for USB external camera.
• img_types: Numbers of object types that you want to include in your dataset.
• label: List of object names to include in the dataset. The constructor creates a named folder for every image type.
• num: Select the number of images you want to shoot and save them in every object folder.
• height: Frame height values.
• width: Frame width values.

How to take pictures manually

In the title bar of the image window, after the caption identifying the image type, there are instructions for manually taking pictures using the ManualDataset. These instructions will be displayed as follows:

• Camera view for image type . Press [s] on the keyboard to save image number: .

---

Build datasets from different npz files (BuildDataset)

If you want to create a dataset that includes different types of images, but cannot shoot all image types at once (e.g., due to time constraints or location differences), you can use the BuildDataset class.

• Create datasets with the just_compress method;
• Create different .npz files for every type of images that composes the dataset (use the same colour spaces in all datasets, RGB or grayscale);
• Create a new script and call the BuildDataset class that merges all the .npz files created before.

Script example

import xeye
# list of directory (paths for the .npz files)
path = ['batch_2.npz', 'batch_3.npz']
# list of labels associated with the images inside the .npz files
label = [0,1]
data = xeye.BuildDataset(path=path, label=label, size = None, color=True, split=True, perc=0.2)
data.build()

The parameters passed to the class BuildDataset:

• path: List of files (.npz) path you want to include in the new dataset
• label: List of ordinal integer values representing the class type of the images inside a .npz file contained in the new dataset. In the example script, the first .npz file images are associated with class 0, while the second .npz file images are associated with class 1. Remember: always start with 0.
• size: Tuple (height, width) for the images in the new dataset created. The default value (None) indicates that new images have the maximum height and width found in the datasets listed as dimensions
• color: Defines if the images contained in the .npz files are RGB or grayscale. A boolean value, by default, is set to True (meaning RGB images).
• split: Defines if you want to build a dataset split in train-test or not. A boolean value, by default, is set to True.
• perc: Defines the percentage of images assigned to the test dataset. A floating value between (0,1). It's set to 0.1 by default.

When you want to use the BuildDataset class, you need to have .npz files containing images with the same types of colour spaces (all grayscale images or RGB).