jasmcaus / caer Goto Github PK

High-performance Vision library in Python. Scale your research, not boilerplate.

License: MIT License

Python 99.24% Shell 0.48% Makefile 0.28%

python vision artificial-intelligence machine-learning neural-network deep-learning gpu ai data-science image-processing

caer's Introduction

Caer - Modern Computer Vision on the Fly

Caer is a lightweight, high-performance Vision library for high-performance AI research. We wrote this framework to simplify your approach towards Computer Vision by abstracting away unnecessary boilerplate code giving you the flexibility to quickly prototype deep learning models and research ideas. The end result is a library quite different in its design, that’s easy to understand, plays well with others, and is a lot of fun to use.

Our elegant, type-checked API and design philosophy makes Caer ideal for students, researchers, hobbyists and even experts in the fields of Deep Learning and Computer Vision.

Overview

Caer is a Python library that consists of the following components:

Component	Description
caer	A lightweight GPU-accelerated Computer Vision library for high-performance AI research
caer.color	Colorspace operations
caer.data	Standard high-quality test images and example data
caer.path	OS-specific path manipulations
caer.preprocessing	Image preprocessing utilities.
caer.transforms	Powerful image transformations and augmentations
caer.video	Video processing utilities

Usually, Caer is used either as:

a replacement for OpenCV to use the power of GPUs.
a Computer Vision research platform that provides maximum flexibility and speed.

Installation

See the Caer Installation guide for detailed installation instructions (including building from source).

Currently, caer supports releases of Python 3.6 onwards; Python 2 is not supported (nor recommended). To install the current release:

$ pip install --upgrade caer

Getting Started

Minimal Example

import caer

# Load a standard 640x427 test image that ships out-of-the-box with caer
sunrise = caer.data.sunrise(rgb=True)

# Resize the image to 400x400 while MAINTAINING aspect ratio
resized = caer.resize(sunrise, target_size=(400,400), preserve_aspect_ratio=True)

For more examples, see the Caer demos or Read the documentation

Resources

Contributing

We appreciate all contributions, feedback and issues. If you plan to contribute new features, utility functions, or extensions to the core, please go through our Contribution Guidelines.

To contribute, start working through the caer codebase, read the Documentation, navigate to the Issues tab and start looking through interesting issues.

Current contributors can be viewed either from the Contributors file or by using the caer.__contributors__ command.

Asking for help

If you have any questions, please:

License

Caer is open-source and released under the MIT License.

BibTeX

If you want to cite the framework feel free to use this (but only if you loved it 😊):

@article{jasmcaus,
  title={Caer},
  author={Dsouza, Jason},
  journal={GitHub. Note: https://github.com/jasmcaus/caer},
  volume={2},
  year={2020-2021}
}

caer's People

Contributors

Stargazers

Watchers

Forkers

fossabot elbert-lau dauriata ekans-tee arevolutioner bhardwajrahul templeblock crowler70 sailfish009 raijinspecial flavio58it srinivasaragavan aaronxiongli goadane luorixiangyang chuckjhao arnold265 malikmalna gitter-badger kmamine oflenake soheibkadi roysti10 avishkar7 tonytaowang emanokun ashokfinds rockyzdx anirbansahaanik rainfallnz vivekreddyk jianfei209 ccf19881030 restevesd avasquez-80 raysaini19 muhammadfadhilarkan roijo tlrlutz22 kyleston fayel-cyber mustafa-sarshar mng222n cflin-cjcu msg4rajesh mustafabozkaya pradiptapradhanang jzialesi thesekyi sunnymakode chendbox sourceexp jameskim-ai coderaayush brccabral edisplay priya1puresoftware arshroop-saini tsis-mobile-technology crossbow-vision ilensb shreyansh2310 reference-repos-dhiman shailesh0209 shahrukh031119 srivats2904 blis1 ishan-phys githubdragonfly ajunlonglive arjunadilan hadryan softwareasimov gg-big-org icodein finden-labs ghimanshu98 vaimalaviya1233 arpitjain799 debikasamanta benkivuva igromanru tennysongeorge sajjadrobin sumsuzzaman tyudihw rifqacodes kbitc koyongje ryandran tchallaalbertkwame kinlead chrischen98 owoeye-babatunde sumitkumar-17 xiaqianseed iamnandini faisalmehmood2007

caer's Issues

Add Type Annotations [OPEN FOR CONTRIBUTIONS]

There are quite a few functions in the repo that need to be annotated.
This issue will be left open for external contributions. If you wish to contribute, please submit a PR for the same!

Thanks!
-J

AttributeError: module 'caer' has no attribute 'normalize'

Hey sorry to bother again. I upgraded caer to 2.0.1
I proceeded but again
from tensorflow.keras.utils import to_categorical
--> featureSet = caer.normalize(featureSet)
labels = to_categorical(labels, len(characters))

the second line throws an error.
AttributeError: module 'caer' has no attribute 'normalize'

[Feat] Add support for the LUV colorspace in `caer.color`

Add support for the LUV colorspace in caer.color

Open for contributions :)

ValueError: verbose flag must be either 1 (display progress to terminal) or 0 otherwise

I am learning OpenCV and at first, I saw the youtube video from freecodecamp OpenCv Course by Jason Dzousa In the last video the line:
train = caer.preprocess_from_dir(char_path, characters, IMG_SIZE, channels, isShuffle=True)
throws the error:
ValueError: verbose flag must be either 1 (display progress to terminal) or 0 otherwise
I can't understand why this error comes.

Enable typing support

[Bug] Tests Failing - no attribute named 'transforms'

Tests failing - assert b/w BGR and RGB

Cannot import BGR2HLS from caer.color

imsave function does not return status

🚀 caer,imsave

For your release, in caer.io.imsave function, you can see that the function does not return any status.

Motivation

See that opencv already return status, we can integrate the status with some personal control.

Updating Project Copyright

📚 Documentation

Hey I noticed that you all still have the copyright for the software set to 2020. Would it be okay for me to go into the codebase and update that for 2021 when you've confirmed that the copyright has been updated, assuming you haven't already done that yet.

Add docs for caer.video

Update docstrings for caer.color

ModuleNotFoundError: No module named 'cv2.cv2'

import caer

ModuleNotFoundError Traceback (most recent call last)
in
----> 1 import caer

~/QWE/ASD/ZXC/FGH/VBN/caer/init.py in
70
71 # Preprocessing
---> 72 from .preprocess import (
73 preprocess_from_dir,
74 sep_train,

~/QWE/ASD/ZXC/FGH/VBN/caer/preprocess.py in
16
17 from .utilities import get_classes_from_dir, to_array
---> 18 from .io import imread
19 from .preprocessing import MeanProcess
20 from ._internal import _check_target_size, _check_mean_sub_values

~/QWE/ASD/ZXC/FGH/VBN/caer/io.py in
11
12
---> 13 import cv2 as cv
14 from urllib.request import urlopen
15

~/QWE/ASD/ZXC/FGH/VBN/cv2/init.py in
3 import sys
4
----> 5 from .cv2 import *
6 from .data import *
7

ModuleNotFoundError: No module named 'cv2.cv2'

caer.preprocess_from_dir() Error: File "C:\Python38\lib\site-packages\caer\preprocess.py

Hi,
I am trying out your code for face recognition using Tensorflow's implementation of Keras. I am having an issue when creating the training data, with the following line:
"train = caer.preprocess_from_dir(...)"

As datasets, I have several folders with 10-16 images in each folder. And in the previous lines of code everything worked fine but when I am trying to train the images, it is processing only 3 images of the first folder (which contains 16 images). It is processing the last 3 images of the first folder, then it crashes.

The error is as follows:
File "C:\Python38\lib\site-packages\caer\preprocess.py", line 132, in preprocess_from_dir for image_path in img_list:
TypeError: 'NoneType' object is not iterable

Thanks

[Feat] Add support for the YUV colorspace

Title ^