matplot RGB vs opencv BGR vs caffe images

Coding/Image

by linguana 2021. 4. 9. 22:06

matplot RGB vs opencv BGR vs caffe images | C++ Python. Computer Vision Deep Learning | KeZunLin's Blog

Guide Matplot (skimage/ PIL Image) import matplotlib.pyplot as plt import matplotlib.image as img image = img.imread("images/cat.jpg") print image.shape print image[:5,:5,0] plt.imshow(image) plt.show() (360, 480, 3) [[26 27 25 28 30] [26 27 25 26 28] [26

kezunlin.me

각 라이브러리마다 이미지 정보를 담는 형식이 다르다.

	dims	order	range	dtype
matplotlib.pyplot	HWC	RGB	[0,255]	uint8
PIL.Image	HWC	RGB	[0,255]	uint8
skimage	HWC	RGB	[0,255]	unit8
OpenCV	HWC	BGR	[0,255]	uint8
caffe.io.load_image	HWC	RGB	[0,1]	float32

※ channel last 로 다 되어있는 이유는 벡엔드를 Theano로 하지 않아서인듯

machinelearningmastery.com/a-gentle-introduction-to-channels-first-and-channels-last-image-formats-for-deep-learning/

A Gentle Introduction to Channels-First and Channels-Last Image Formats

Color images have height, width, and color channel dimensions. When represented as three-dimensional arrays, the channel dimension for the image data is last by default, but may be moved to be the first dimension, often for performance-tuning reasons. The

machinelearningmastery.com

백엔드에 따라서 Channel 순서가 달라진다.

TensorFlow: Channels last order.
Theano: Channels first order.
CNTK: Channels last order.

저작자표시 비영리 변경금지 (새창열림)

'Coding > Image' 카테고리의 다른 글

[2] Selective Search (0)	2021.04.12
[1] CNN Image Classifier to Object Detector (0)	2021.04.12
OpenCV fundamentals (0)	2021.04.09
Scene Text Recognition (0)	2021.04.01
Object Detection (R-CNN) (0)	2021.03.26