Statistical Classification of Image Content for Visual Information Filtering
نویسندگان
چکیده
An increasing number of freely accessible adult content websites arose recently, displaying a wide variety of different offensive images and videos. Since many users do not want to be confronted with such material, automatic tools to detect and filter these images and videos are needed. Additionally, tools are required to protect children from accessing offensive websites. This thesis presents approaches for both classification of offensive images and videos. For the first, two different approaches are presented and evaluated on a variety of different datasets showing real world Web content. One traditional method is based on detecting and describing skin areas, while the other uses the popular bag-of-visual-words model. Video classification is based on keyframes and additional motion features, including periodicity detection. Evaluation of these techniques is done on offensive Web videos and inoffensive YouTube videos. The results show that the bag-of-visual-words approach is better suited for classifying offensive material than traditional skin features. Also, combining keyframe classification with additional motion features improves the performance of detecting offensive videos. Overall, a classification accuracy of 99% on images, and 94% on videos is reached.
منابع مشابه
Palarimetric Synthetic Aperture Radar Image Classification using Bag of Visual Words Algorithm
Land cover is defined as the physical material of the surface of the earth, including different vegetation covers, bare soil, water surface, various urban areas, etc. Land cover and its changes are very important and influential on the Earth and life of living organisms, especially human beings. Land cover change monitoring is important for protecting the ecosystem, forests, farmland, open spac...
متن کاملImage spam filtering using textual and visual information
In this paper we focus on the so-called image spam, which consists in embedding the spam message into images attached to e-mails to circumvent statistical techniques based on the analysis of body text of e-mails (like the “bayesian filters”), and in applying content obscuring techniques to such images to make them unreadable by standard OCR systems without compromising human readability. We arg...
متن کاملAn Adaptive Hierarchical Method Based on Wavelet and Adaptive Filtering for MRI Denoising
MRI is one of the most powerful techniques to study the internal structure of the body. MRI image quality is affected by various noises. Noises in MRI are usually thermal and mainly due to the motion of charged particles in the coil. Noise in MRI images also cause a limitation in the study of visual images as well as computer analysis of the images. In this paper, first, it is proved that proba...
متن کاملSteganography Scheme Based on Reed-Muller Code with Improving Payload and Ability to Retrieval of Destroyed Data for Digital Images
In this paper, a new steganography scheme with high embedding payload and good visual quality is presented. Before embedding process, secret information is encoded as block using Reed-Muller error correction code. After data encoding and embedding into the low-order bits of host image, modulus function is used to increase visual quality of stego image. Since the proposed method is able to embed...
متن کاملImage Classification via Sparse Representation and Subspace Alignment
Image representation is a crucial problem in image processing where there exist many low-level representations of image, i.e., SIFT, HOG and so on. But there is a missing link across low-level and high-level semantic representations. In fact, traditional machine learning approaches, e.g., non-negative matrix factorization, sparse representation and principle component analysis are employed to d...
متن کامل