This paper investigates cellular automaton (CA)-based segmentation techniques on microscopic images. An image is considered a 2D cellular automaton, and the pixel values are considered the states of the cells. For updating the state of the cell, von Neumann and Moore neighbourhoods have been employed. An existing dataset of microscopic images has been taken, and various segmentation rules have been used for analysis. Additionally, a comparison has been made among the methods, with results demonstrating satisfactory performance. It is also observed that the von Neumann neighbourhood-based approach is producing better results compared to the Moore one.

Reachability tree proves its power to characterize the elementary cel-lular automata under logic-0 constant (null) boundary condition and periodic boundary condition. A reachability tree implicitly represents the configuration space of a cellular automaton (CA). It reveals various aspects of CA, such as identification of reachable or non-reachable con -figuration, cyclic or acyclic configurations, determining reversibility or irreversibility of a given CA, and so on. This paper contributes towards the characterization of cellular automata under the open boundary con-ditions. A generalized view of reachability tree is reported, which tar -gets the different categories of open boundary condition -adiabatic boundary condition, reflexive boundary condition and constant bound-ary condition (logic-1 & logic-0). This work establishes that identical CA configuration space can be generated using cellular automata under the above mentioned boundary conditions, for a finite sized CA. This work also reports the characterization of reachability tree for CA under intermediate boundary condition which is another category of open boundary condition cellular automata.

Due to the difficulty of the problem, there have been many different approaches and algorithms developed to date. Defocus deblurring, single image motion deblurring, image deraining, and image denoising are some of the further branches of picture restoration. There are numerous algorithms that are appropriate in a variety of circumstances. For instance, gaussian noise like pepper and salt noise responds well to the median filter. The adaptive filter, on the other hand, performs particularly well in DSP and ANC applications. It has been discovered that methods based on deep neural networks, such as the restormer, can occasionally outperform earlier algorithms. We look for the best algorithm for image denoising, defocus deblurring, single image motion deblurring, and picture deraining in this work. Although restormer is shown to perform better than the others in the majority of circumstances, hybrid median filter is sometimes found to perform better than all of them combined.

Naskar, Nazma Sivaraj, B. K.In this paper the elementary cellular automata is explored under adiabatic boundary condition and reflexive boundary condition, using the reachability tree. To construct the reachability tree under the mentioned conditions, structures of those boundary conditions are analysed. The valid RMTs of root level and leaf level are identified. It was found that only at root level and leaf level, the structures and arrangements differ slightly with the reachability tree of null boundary condition.

This paper will contribute mainly to two aspects: (i) A synthesis scheme of cellular automata having only point attractor, for specified pseudo-exhaustive bit (PE-bit), under periodic boundary condition, and (ii) a designing classification scheme for imbalance dataset. In his paper, another parallel objective will be carried out. That is, study whether the boundary conditions affect the performance of the classifier? To design classifier researchers are already shown that cellular automata having only point attractor are efficiently used to design a classifier. In this paper, cellular automata having only point attractor for specified pseudo-exhaustive bit will be used to design classifiers and both the null and periodic boundary conditions will be considered.

In this work, we study and analyze different feature selection algorithms that can be used to classify cancer subtypes in case of highly varying high-dimensional data. We apply three different feature selection methods on five different types of cancers having two separate omics each. We show that the existing feature selection methods are computationally expensive when applied individually. Instead, we apply these algorithms sequentially which helps in lowering the computational cost and improving the predictive performance. We further show that reducing the number of features using some dimension reduction techniques can improve the performance of machine learning models in some cases. We support our findings through comprehensive data analysis and visualization.

Cellular automata (CAs) are dynamical systems which exhibit complex global behavior from simple local interaction and computation. Since the inception of cellular automaton (CA) by von Neumann in 1950s, it has attracted the attention of several researchers over various backgrounds and fields for modeling different physical, natural as well as real-life phenomena. Classically, CAs are uniform. However, non-uniformity has also been introduced in update pattern, lattice structure, neighborhood dependency and local rule. In this survey, we tour to the various types of CAs introduced till date, the different characterization tools, the global behavior of CAs, like universality, reversibility, dynamics etc. Special attention is given to non-uniformity in CAs and especially to non-uniform elementary CAs, which have been very useful in solving several real-life problems.

This paper presents an efficient scheme of determining cycle structure of a finite one dimensional non-uniform cellular automata (CA). We process the reachability tree of the CA and retain only those edges which are involved in cycles. Our scheme reduces both the space and time complexity, since only those rule min terms involved in crosslinks or self links, are considered at each level.

This paper presents the concept of multi-class classmer using non-uniform cellular Automata under periodic boundary condition. To design this multi-classifier, a special class of non-uniform cellular automata (CA) that contain point attractors in their state space, is studied. These CA always converge to some point attractors. The reachability tree, a discrete tool for characterizing l-d CA and Link, have been utilized to develop theories for these types of CA. We report an algorithm that synthesizes a non-uniform cellular automaton having only point attracto/rs.

This paper studies a special class of non-uniform cellular automata (CAs) that contain only single length cycle (point) attractors in their state space. These CAs always converge to some point attractors. A number of theorems and lemmas are reported in this paper to characterize this class of CAs. Reachability tree, a discrete tool for characterizing 1-d CA, has been utilized to develop theories for these types of CAs. We finally report an algorithm that synthesizes a non-uniform cellular automaton having only point attractors.

This paper reports a scheme to identify cyclic/acyclic states from the state space of cellular automata (CA). We analyze the reachability tree to do this. An algorithm is presented to count the cyclic states. We introduce a concept of tree merging for identifying cyclic and acyclic states. To our knowledge, this is the first work to efficiently identify cyclic/acyclic states from the state space of nonlinear CA. Since cyclic states can only form attractors, the identification of cyclic states would help us to characterize CA attractors.

Characterization of a special class of cellular automata (CA) having only point attractors is the necessity to devise CA based solutions for diverse applications, like pattern classification, pattern recognition, etc. This work explores the essential properties of attractors towards characterization of the 1-dimensional nonlinear CA with point attractors (single length cycle attractors). The concept of Reachability Tree is introduced for such characterization. It enables identification of the pseudo-exhaustive bits (PE-bits) of a CA defining the point attractors. The theoretical framework, developed, is found to be most effective to devise schemes for synthesizing a single length cycle multiple attractor CA, in linear time, for a given set of attractors with the specific set of PE-bits. This finally enables classification of CA rules that can be employed for synthesizing an n-cell CA, for an arbitrary a, having single length cycle attractors only.

The single attractor cellular automata (SACA) is of prime interest in devising schemes for different applications specially in authentication and cryptography. The synthesis of SACA in linear/additive domain has been proposed in literature. This work reports characterization of such a special class of CA beyond linear domain. The characterization is based on the analysis of individual CA rule and its potential to form the single length cycle attractors (point states). The proposed characterization targets design of a CA based scheme for detection of faulty nodes in a wireless sensor network. It enables identification of faults even in multiple nodes with out major computation overhead.

The special class of irreversible cellular automaton (CA) with multiple attractors is of immense interest to the CA researchers. Characterization of such a CA is the necessity to devise CA based solutions for diverse applications. This work explores the essential properties of CA attractors towards characterization of the 1-dimensional cellular automata with point states (single length cycle attractors). The concept of Reachability Tree is introduced for such characterization. It enables identification of the pseudo-exhaustive bits (PE bits) of a CA defining its point states. A theoretical framework has been developed to devise schemes for synthesizing a single length cycle multiple attractor CA with the specific set of PE bits. It also results in a linear time solution while synthesizing a CA for the given set of attractors and its PE bits. The experimentation establishes that the proposed CA synthesis scheme is most effective in designing the efficient pattern classifiers for wide range of applications.

This work proposes characterization of single-length cycle cellular automata (CA) attractors with the target to model this class of CA for designing efficient pattern recognizer. Identification of essential properties of a CA while forming multi-length cycles provides the basis of such characterization. A scheme has been developed that synthesizes the single-length cycle attractor CA, avoiding multi-length cycles, desired for a pattern recognizer.