This paper presents a patch-based inpainting algorithm for image block recovery in block-based coding image transmission. The algorithm is based on a geometric model for patch synthesis. The lost pixels are recovered by copying pixel values from the source using a similarity criterion. We used a trained neural network to choose the “best similar” patch. Experimental results show that the proposed method outperforms widely used state-of-the-art methods in both subjective and objective measurements of image block recovery.
Due to an impressive progress in digital imaging reached over past decades, the importance of analog video systems as primary instruments of receiving, transmitting and storing data has been greatly reduced. However, there still exist large amount of data stored in analog formats on media affected by aging. In this paper, a new three-stage method for detecting and restoring blotches on a video sequence has been developed. The new method including the motion compensation, LBP calculations and data classification using the neural network is shown to have higher efficiency than commonly used ROD, SROD and SDI methods.
Digital video forgery or manipulation is a modification of the digital video for fabrication, which includes frame sequence manipulations such as deleting, insertion and swapping. In this paper, we focus on the detection problem of deleted frames in videos. Frame dropping is a type of video manipulation where consecutive frames are deleted to skip content from the original video. The automatic detection of deleted frames is a challenging task in digital video forensics. This paper describes an approach using spatial-temporal analysis based on the convolution with a bank of 3D Gabor filters. Also, we use the 3D Convolutional Neural Network for frame drop detection for preprocessed frames. Experimental results demonstrate the effectiveness of the proposed approach on a test video database.
This paper describes a framework for action recognition which aims to recognize the goals and activities of one or more human from a series of observations. We propose an approach for the human action recognition based on the 3D dense micro-block difference. The proposed algorithm is a two-stage procedure: (a) image preprocessing using a 3D Gabor filter and (b) a descriptor calculation using 3D dense micro-block difference with SVM classifier. At the first step, an efficient spatial computational scheme designed for the convolution with a bank of 3D Gabor filters is present. This filter intensifies motion using a convolution for a set of 3D patches and arbitrarily-oriented anisotropic Gaussian. For preprocessed frames, we calculate the local features such as 3D dense micro-block difference (3D DMD), which capture the local structure from the image patches at high scales. This approach is processing the small 3D blocks with different scales from frames which capture the microstructure from it. The proposed image representation is combined with fisher vector method and linear SVM classifier. We evaluate the proposed approach on the UCF50, HMDB51 and UCF101 databases. Experimental results demonstrate the effectiveness of the proposed approach on video with a stochastic textures background with comparisons of the state-of-the-art methods.