{"cells":[{"metadata":{},"cell_type":"markdown","source":"# LEts Start First with What Is Steganography ?\nSteganography is to conceal the secret data within multimedia contents such as file, message, image, or video. Steganography is concerned with concealing the fact that the secret data is being sent covertly as well as concealing the contents of the secret data . Steganalysis is the counter part of steganography that defined as the art of science of detecting the hidden secret data in cover objects. In other words, steganalysis is to detect secret data hidden using steganography, where steganalysis is to identify suspected packages, determine whether the secret data is embedded or not . Machine learning is a field of artificial intelligence to provide the ability to learn without being programmed and deep learning is a subset of machine learning . In this paper, steganalysis and machine learning techniques are explained and the process and possibility for steganalysis in various machine learning frameworks are described. Some datasets on stego-images are prepared and training model are tested. "},{"metadata":{},"cell_type":"markdown","source":"# What is Actually Stegnanalysis ?\n\nStegnanalysis is the art of secret communication and steganalysis is the art of detecting the hidden messages embedded in digital media using steganography. Both steganography and steganalysis have received a great deal of attention from law enforcement and the media. In the past years many powerful and robust methods of steganography and steganalysis have been reported in the literature. In this paper, we classify and give an account of the various approaches that have been proposed for steganalysis. Some promising methods for statistical steganalysis have also been identified."},{"metadata":{},"cell_type":"markdown","source":"\n","attachments":{}},{"metadata":{},"cell_type":"markdown","source":"**Steganalysis** is to identify suspected data, determine hidden data, and recover the hidden data. Steganalysis can be divided into **four** categories: \n- visual \n- structural \n- statistical \n- learning steganalysis. \n\n**Visual steganalysis** is to investigate visual artifacts in the stego-images, where try to catch visual difference by analyzing stego-images. \n\n**Structural steganalysis** looks into suspected signs in the media format representation since the format is often changed when the secret message is embedded. RS analysis and pair analysis are included in the structural steganalysis.\n\n**Statistical steganalysis** utilizes statistical models to detect steganography techniques. Statistical steganalysis can be divided into specific statistical and universal statistical steganalysis.\n\n**Learning steganalysis** also called blind steganalysis is one of universal statistical steganalysis since cover images and stego-images are used as training datasets. Other classification of steganalysis can be divided into six categories . It is depending on what kind of attacks a forensic examiner uses. "},{"metadata":{},"cell_type":"markdown","source":"# EDA Stegnanalysis"},{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"import numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nimport imageio\nimport matplotlib.pyplot as plt\nimport matplotlib.image as mpimg\nimport torch\n","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"cell_type":"code","source":"PATH = '/kaggle/input/alaska2-image-steganalysis'\nmodel = '/kaggle/input/alaska2-image-steganalysis/e4s-srm/'\n\ntrain_images = pd.Series(os.listdir(PATH + '/Cover/')).sort_values(ascending=True).reset_index(drop=True)\ntest_images = pd.Series(os.listdir(PATH + '/Test')).sort_values(ascending=True).reset_index(drop=True)\nsample_submission = pd.read_csv(f'{PATH}/sample_submission.csv')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"es4 = .alexnet(pretrained=True)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"sample_submission.head()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Display Images"},{"metadata":{"trusted":true},"cell_type":"code","source":"fig, ax = plt.subplots(nrows=2, ncols=4, figsize=(16, 8))\nk=0\nfor i, row in enumerate(ax):\n for j, col in enumerate(row):\n img = mpimg.imread(PATH + '/Cover/' + train_images[k])\n col.imshow(img)\n col.set_title(train_images[k])\n k=k+1\nplt.suptitle('Samples from Cover Images', fontsize=14)\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"fig, ax = plt.subplots(nrows=2, ncols=4, figsize=(16, 8))\nk=0\nfor i, row in enumerate(ax):\n for j, col in enumerate(row):\n img = mpimg.imread(PATH + '/Test/' + test_images[k])\n col.imshow(img)\n col.set_title(test_images[k])\n k=k+1\nplt.suptitle('Samples from Test Images', fontsize=14)\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"for folder in os.listdir(PATH):\n try:\n print(f\"Folder {folder} contains {len(os.listdir(PATH + '/' + folder))} images.\")\n except:\n print(f'{folder}')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Displaying JMiPOD , JUNIWARD , UERD, with train"},{"metadata":{"trusted":true},"cell_type":"code","source":"folders = ['Cover', 'JMiPOD', 'JUNIWARD', 'UERD']\nk=0\nimg_cover = mpimg.imread(PATH + '/' + folders[0] + '/'+ train_images[k])\nimg_jmipod = mpimg.imread(PATH + '/' + folders[1] + '/'+ train_images[k])\nimg_juniward = mpimg.imread(PATH + '/' + folders[2] + '/'+ train_images[k])\nimg_uerd = mpimg.imread(PATH + '/' + folders[3] + '/'+ train_images[k])\n\nfig, ax = plt.subplots(nrows=3, ncols=4, figsize=(16, 12))\n\nax[0,0].imshow(img_jmipod)\nax[0,1].imshow((img_cover == img_jmipod).astype(int)[:,:,0])\nax[0,1].set_title(f'{train_images[k]} Channel 0')\n\nax[0,2].imshow((img_cover == img_jmipod).astype(int)[:,:,1])\nax[0,2].set_title(f'{train_images[k]} Channel 1')\nax[0,3].imshow((img_cover == img_jmipod).astype(int)[:,:,2])\nax[0,3].set_title(f'{train_images[k]} Channel 2')\nax[0,0].set_ylabel(folders[1], rotation=90, size='large', fontsize=14)\n\n\nax[1,0].imshow(img_juniward)\nax[1,1].imshow((img_cover == img_juniward).astype(int)[:,:,0])\nax[1,2].imshow((img_cover == img_juniward).astype(int)[:,:,1])\nax[1,3].imshow((img_cover == img_juniward).astype(int)[:,:,2])\nax[1,0].set_ylabel(folders[2], rotation=90, size='large', fontsize=14)\n\nax[2,0].imshow(img_uerd)\nax[2,1].imshow((img_cover == img_uerd).astype(int)[:,:,0])\nax[2,2].imshow((img_cover == img_uerd).astype(int)[:,:,1])\nax[2,3].imshow((img_cover == img_uerd).astype(int)[:,:,2])\nax[2,0].set_ylabel(folders[3], rotation=90, size='large', fontsize=14)\n\nplt.suptitle('Pixel Deviation from Cover Image', fontsize=14)\n\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Total number of images"},{"metadata":{"trusted":true},"cell_type":"code","source":"print('{} images with Cover Images '.format(train_images.nunique()))\nprint('{} images with Test Images '.format(test_images.nunique()))\n\n","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat":4,"nbformat_minor":4}