Volume 10 (2020): Issue 1 (January 2020)

The on-line signature is a biometric attribute which can be used for identity verification. It is a very useful characteristic because it is commonly accepted in societies across the world. However, the verification process using this particular biometric feature is a rather difficult one. Researchers working on identity verification involving the on-line signature might face various problems, including the different discriminative power of signature descriptors, the problem of a large number of descriptors, the problem of descriptor generation, etc. However, population-based algorithms (PBAs) can prove very useful when resolving these problems. Hence, we propose a new method for on-line signature partitioning using a PBA in order to improve the verification process effectiveness. Our method uses the Differential Evolution algorithm with a properly defined evaluation function for creating the most characteristic partitions of the dynamic signature. We present simulation results of the proposed method for the BioSecure DS2 database distributed by the BioSecure Association.

In recent years, many deep learning methods, allowed for a significant improvement of systems based on artificial intelligence methods. Their effectiveness results from an ability to analyze large labeled datasets. The price for such high accuracy is the long training time, necessary to process such large amounts of data. On the other hand, along with the increase in the number of collected data, the field of data stream analysis was developed. It enables to process data immediately, with no need to store them. In this work, we decided to take advantage of the benefits of data streaming in order to accelerate the training of deep neural networks. The work includes an analysis of two approaches to network learning, presented on the background of traditional stochastic and batch-based methods.

In this paper we tackle the problem of vehicle re-identification in a camera network utilizing triplet embeddings. Re-identification is the problem of matching appearances of objects across different cameras. With the proliferation of surveillance cameras enabling smart and safer cities, there is an ever-increasing need to re-identify vehicles across cameras. Typical challenges arising in smart city scenarios include variations of viewpoints, illumination and self occlusions. Most successful approaches for re-identification involve (deep) learning an embedding space such that the vehicles of same identities are projected closer to one another, compared to the vehicles representing different identities. Popular loss functions for learning an embedding (space) include contrastive or triplet loss. In this paper we provide an extensive evaluation of triplet loss applied to vehicle re-identification and demonstrate that using the recently proposed sampling approaches for mining informative data points outperform most of the existing state-of-the-art approaches for vehicle re-identification. Compared to most existing state-of-the-art approaches, our approach is simpler and more straightforward for training utilizing only identity-level annotations, along with one of the smallest published embedding dimensions for efficient inference. Furthermore in this work we introduce a formal evaluation of a triplet sampling variant (batch sample) into the re-identification literature. In addition to the conference version [24], this submission adds extensive experiments on new released datasets, cross domain evaluations and ablation studies.

The paper presents the idea of connecting the concepts of the Vapnik’s support vector machine with Pawlak’s rough sets in one classification scheme. The hybrid system will be applied to classifying data in the form of intervals and with missing values [1]. Both situations will be treated as a cause of dividing input space into equivalence classes. Then, the SVM procedure will lead to a classification of input data into rough sets of the desired classes, i.e. to their positive, boundary or negative regions. Such a form of answer is also called a three–way decision. The proposed solution will be tested using several popular benchmarks.

Fast content-based image retrieval is still a challenge for computer systems. We present a novel method aimed at classifying images by fuzzy rules and local image features. The fuzzy rule base is generated in the first stage by a boosting procedure. Boosting meta-learning is used to find the most representative local features. We briefly explore the utilization of metaheuristic algorithms for the various tasks of fuzzy systems optimization. We also provide a comprehensive description of the current best-performing DISH algorithm, which represents a powerful version of the differential evolution algorithm with effective embedded mechanisms for stronger exploration and preservation of the population diversity, designed for higher dimensional and complex optimization tasks. The algorithm is used to fine-tune the fuzzy rule base. The fuzzy rules can also be used to create a database index to retrieve images similar to the query image fast. The proposed approach is tested on a state-of-the-art image dataset and compared with the bag-of-features image representation model combined with the Support Vector Machine classification. The novel method gives a better classification accuracy, and the time of the training and testing process is significantly shorter.