kenneth111 / pyimageprocess

A Python library to process YUV files and images

PyImageProcess

A Python library to process images and YUV files.

Dependencies:

• numpy
• PIL
• scipy

utils.py

get_random_pics(img, h, w, num)

This function is developed for training a deep neural network. It can extract a number of patches from a picture. In addtion, this function is a generator and can be used in a for loop.

img: an Image object (PIL.Image)

h: the height of a patch

w : the width of a patch

num: how many patches to extract

get_batch_pics(img, h, w, num)

This function is developed for training a deep neural network. It extracts a number of patches from a picture and pack the patches into a ndarray object. This function calls get_random_pics to get random patches.

img: an Image object

h: the height of a patch

w: the width of a patch

num: how many patches to extract

return: a 4 dimension ndarray object: [num, w, h, 3]

get_all_pics(img, h, w)

This function is developed for training a deep neural network. It extracts (height / h) * (width / w) patches from a picture and these patches are not overlapped. In addtion, this function is a generator and can be used in a for loop.

img: an Image object (PIL.Image)

h: the height of a patch

w : the width of a patch

utils_yuv.py

get_a_frame(filename, height, width, frameId)

Extract an frame from a YUV file. This function returns an n by 3 matrix containing Y, U and V components.

filename: a string, like "capture.yuv"

height: the frame's height

width: the frame's width

frameId: the frame number to be extracted, starting from 1

get_all_frames(filename, height, width, num_frames)

Extract all the frames and their corrsponding frame ID in a YUV file. This function is a generator and therefore can be used in a for loop.

filename: a string, like "capture.yuv"

height: the frame's height

width: the frame's width

num_frames: how many frames in the YUV file

yuv_to_rgb(yuv_frame)

Convert YUV to RGB . The conversion formula can be found here.

yuv_frame: an n by 3 matrix containing Y, U and V seperately.

return: an n by 3 matrix containing R, G and B seperately.

def convertYUV2Img(yuv_filename, height, width, num_frames, format, img_filename)

Read frames from a yuv file and save the frame to image files

yuv_filename: the yuv file to convert

height, width: the height and the width of the yuv file

num_frames: how many frames of the yuv file to convert

format: such as "jpg", "bmp"

img_filename: a string used in image files

yuv_draw_box(yuv_frame, height, width, startx, starty, endx, endy, color)

Draw a box on a YUV frame.

yuv_frame: a n by 3 matrix containing Y, U and V, seperately, or a (height, width, 3) matrix

color: "r", "g" or "b"

utils_rgb.py

show_img_diff(file1, file2, diff_file)

Compare two images and save the difference between these images into a new image file.

file1: the first image file to compare (string)

file2: the second image file to compare (string)

diff_file: where to save the difference (string)

rgb_to_yuv(rgb_frame)

Convert RGB to YUV. The conversion formula can be found here.

rgb_frame: an n by 3 matrix containing R, G, and B seperately.

return: an n by 3 matrix containing Y, U and V seperately.

save_rgb(filename, rgb, height, width)

Save a RGB frame to a file.

filename: The file name, like "test.bmp".

rgb: a n by 3 matrix containing R, G and B seperately.

transform.py

dct2(a)

2-d discrete cosine transform. This function can be found here.

a: 2-d ndarray object

idct2(a)

2-d inverse discrete cosine transform. This function can be found here.

a: 2-d ndarray object

pyimageprocess's Issues

ValueError: all the input array dimensions for the concatenation axis must match exactly, but along dimension 1, the array at index 0 has size 921600 and the array at index 2 has size 0

I tried to run convertYUV2Img('../image', 720, 1280, 1, 'bmp', 'test')