Result after Code12.py

[m-kashani]

TO DO

• [x] Crop the image based on it's bounding box: https://www.kaggle.com/whizzkid/crop-images-using-bounding-box

• [x] https://www.geeksforgeeks.org/python-pil-image-crop-method/

  ---

Labels:

['Past' 'Gorgonia' 'SeaRods' 'Antillo' 'Fish' 'Ssid' 'Orb' 'Other_Coral' 'Apalm' 'Galaxaura'] Image Type: 8 = bits | (2704, 1524) = shape | JPEG = imageType Number of Images: 133

Add an image or two images of each of this objects.

• [x] Past

• [x] Gorgonia

• [x] SeaRods

• [x] Antillo -> Antillogorgia.

• [x] Fish

• [x] Ssid

• [x] Orb

• [x] Other_Coral

• [x] Apalm

• [x] Galaxaura https://en.wikipedia.org/wiki/Galaxaura

How many Object in each image:

A_3D_L0646_144.jpg 56 3D_L0622_176.jpg 55 3R010215_829.jpg 54 B_3D_L0647_42.jpg 53 3D_L0622_139.jpg 53 .. 3D_L0457_173.jpg 8 3D_L0443_36.jpg 5 3D_L0441_41.jpg 4 3D_L0453_119.jpg 2 3D_L0457_97.jpg 2

list of top 10:['A_3D_L0646_144.jpg','3D_L0622_176.jpg','3R010215_829.jpg','B_3D_L0647_42.jpg','3D_L0622_139.jpg']

• [x] Added parts of code to generate top20.csv including the images with most amount of objects.

  To do:

• [x] show them all in a clean jupyter notebook.

Objects:

Galaxaura 1151 Gorgonia 880 Antillo 671 Past 326 Apalm 274 Fish 270 SeaRods 221 Orb 118 Other_Coral 58 Ssid 30 Name: image, Length: 133, dtype: int64

To do:

• [x] Plot a histogram of this information. (Auto generator code is preferred)

https://mode.com/python-tutorial/counting-and-plotting-in-python/

[m-kashani]

Showing that the train and test folders are scattered enough: