Abstract:In this paper we propose a novel algorithm that enables online actions segmentation and classification. The algorithm enables segmentation from an incoming motion capture (MoCap) data stream, sport (or karate) movement sequences that are later processed by classification algorithm. The segmentation is based on Gesture Description Language classifier that is trained with an unsupervised learning algorithm. The classification is performed by continuous density forward-only hidden Markov models (HMM) classifier. Our methodology was evaluated on a unique dataset consisting of MoCap recordings of six Oyama karate martial artists including multiple champion of Kumite Knockdown Oyama karate. The dataset consists of 10 classes of actions and included dynamic actions of stands, kicks and blocking techniques. Total number of samples was 1236. We have examined several HMM classifiers with various number of hidden states and also Gaussian mixture model (GMM) classifier to empirically find the best setup of the proposed method in our dataset. We have used leave-one-out cross validation. The recognition rate of our methodology differs between karate techniques and is in the range of 81% ± 15% even to 100%. Our method is not limited for this class of actions but can be easily adapted to any other MoCap-based actions. The description of our approach and its evaluation are the main contributions of this paper. The results presented in this paper are effects of pioneering research on online karate action classification.

In this paper we will experimentally prove that after recalculating the motion capture (MoCap) data to position-invariant representation it can be directly used by classifier to successfully recognize various actions types. The assumption on classifier is that it is capable to deal with objects that are described by hundreds of numeric values. The second novelty of this paper is application of neural network trained with the parallel stochastic gradient descent, Random Forests and Support Vector Machine with Gaussian radial basis kernel to perform classification task on gym exercises and karate techniques MoCap datasets. We have tested our approach on two datasets using k-fold cross-validation method. Depending of the dataset we have obtained averaged recognition rate from 100 to 97 %. Our results presented in this work give very important hints for developing similar actions recognition systems because proposed features selection and classification setup seems to guarantee high efficiency and effectiveness.

This paper presents the new concepts of multisecret and false digital image steganography. The main idea of such approaches is to embed in a single container (digital image) more than one message. The hidden secrets are called real and false messages, respectively. The first one contains essential data which are intended to be securely transferred between different parties, the latter is a bait for focusing attention on an unimportant message. This false and multi-secret steganography will be broken when existence of the real message is revealed, it does not matter whether the false message is detected. Such concepts may find many different applications, especially in situations where communication channel between a sender and a receiver is closely monitored and the warden suspects that the steganography is used. In that case it is probable that the transmitted data will be analyzed in a very detailed way. The concepts described in this paper can help to overcome this problem by dropping a fabricated message and thereby deceiving the warden. The possibility of sending both real and false information at the same time can be seen as additional benefit. In fact, the presented idea allows to establish a kind of subliminal channel while transferring hidden information using digital images. Communicated by V. Loia.

Imagechain is a cryptographic structure that chain digital images with hash links. The most important feature, which differentiates it from blockchain, is that the pictures are not stored inside the blocks. Instead, the block and the image are combined together in the embedding process. Therefore, the imagechain is built from standard graphic files that may be used in the same way as any other image, but additionally, each of them contains a data block that links it to a previous element of the chain. The presented solution does not require any additional files except the images themselves. It supports multiple file formats and embedding methods, which makes it portable and user-friendly. At the same time, the scheme provides a high level of security and resistance to forgery. This is achieved by hashing the whole file with embedded data, so the image cannot be altered or removed from the chain without losing integrity. This article describes the basic concept of an imagechain together with building blocks and applications. The two most important issues are embedding methods and block structure.

No abstract