Classification of indecent videos by low complexity repetitive motion detection

Acoustic scene classification (ASC) categorizes an audio file based on the environment in which it has been recorded. This has long been studied in the detection and classification of acoustic scenes and events (DCASE). This presents the solution to Task 1 of the DCASE 2020 challenge submitted by the Chung-Ang University team. Task 1 addressed two challenges that ASC faces in real-world applications. One is that the audio recorded using different recording devices should be classified in general, and the other is that the model used should have low-complexity. We proposed two models to overcome the aforementioned problems. First, a more general classification model was proposed by combining the harmonic-percussive source separation (HPSS) and deltas-deltadeltas features with four different models. Second, using the same feature, depthwise separable convolution was applied to the Convolutional layer to develop a low-complexity model. Moreover, using gradient-weight class activation mapping (Grad-CAM), we investigated what part of the feature our model sees and identifies. Our proposed system ranked 9th and 7th in the competition for these two subtasks, respectively.

Download Full-text

MobiDB-lite 3.0: fast consensus annotation of intrinsic disorder flavors in proteins

Bioinformatics ◽

10.1093/bioinformatics/btaa1045 ◽

2020 ◽

Author(s):

Marco Necci ◽

Damiano Piovesan ◽

Damiano Clementel ◽

Zsuzsanna Dosztányi ◽

Silvio C E Tosatto

Keyword(s):

Large Scale ◽

Theoretical Models ◽

Intrinsic Disorder ◽

Low Complexity ◽

Compositional Bias ◽

Intrinsically Disordered ◽

Intrinsically Disordered Regions ◽

Proteome Level ◽

And Performance

Abstract Motivation The earlier version of MobiDB-lite is currently used in large-scale proteome annotation platforms to detect intrinsic disorder. However, new theoretical models allow for the classification of intrinsically disordered regions into subtypes from sequence features associated with specific polymeric properties or compositional bias. Results MobiDB-lite 3.0 maintains its previous speed and performance but also provides a finer classification of disorder by identifying regions with characteristics of polyolyampholytes, positive or negative polyelectrolytes, low-complexity regions or enriched in cysteine, proline or glycine or polar residues. Subregions are abundantly detected in IDRs of the human proteome. The new version of MobiDB-lite represents a new step for the proteome level analysis of protein disorder. Availability and implementation Both the MobiDB-lite 3.0 source code and a docker container are available from the GitHub repository:https://github.com/BioComputingUP/MobiDB-lite

Download Full-text

Industrial Verification of Piezo Motors on a CubeSat Based Verification Platform

CANEUS2006: MNT for Aerospace Applications ◽

10.1115/caneus2006-11084 ◽

2006 ◽

Author(s):

Manuel Czech ◽

Ulrich Walter

Keyword(s):

Complex Systems ◽

Low Complexity ◽

Micro Electro Mechanical System ◽

Mission Design ◽

Technology Readiness ◽

Key Technology ◽

Mems Technology ◽

Segmented Mirror ◽

Mirror Telescope

Due to the classification of technologies in NASA’s and ESA’s technology readiness levels, newly developed components have to be space proven before they can be utilized in space missions. This space prove can be adduced by sending these technologies to orbit either as experiment on a piggyback flight or a dedicated mission. Over the last years the size of technologies and satellites has shifted to much smaller sizes. In this paper, the possibility of industrial verification of MEMS (Micro Electro Mechanical System) applications using dedicated pico-satellite missions is examined. Based on the CubeSat concept, a technology verification platform can be realized for verification of not only pico-satellite components, but also of components of complex systems and missions. Therefore a platform fulfilling the requirements for such industrial verification of components named MOVE (Munich Orbital Verification Experiment) is developed at the Institute of Astronautics (LRT). This platform enables professional verification of MEMS technology and techniques at overall mission costs of less than 100k€. As a first application of this approach, a mission called π-MOVE (π for piezo) will verify piezo motors on the developed platform. These piezo motors are representative for components of complex systems, as this motor concept is considered to be key technology for future segmented mirror telescope missions. In the mission design process for this platform, strong emphasis is put on the robustness of the design, low complexity and realizability within the institute’s environment. The advantages through access to both university and industry resources will be taken. The feasibility of professional technology verification is highly dependent on the test plans, which are developed in cooperation with the experienced industrial partners.

Download Full-text

A comparative study of motion detection with FMG and sEMG methods for assistive applications

Journal of Rehabilitation and Assistive Technologies Engineering ◽

10.1177/2055668320938588 ◽

2020 ◽

Vol 7 ◽

pp. 205566832093858

Author(s):

Muhammad Raza Ul Islam ◽

Asim Waris ◽

Ernest Nlandu Kamavuako ◽

Shaoping Bai

Keyword(s):

Motion Detection ◽

Classification Accuracy ◽

Detection Method ◽

Daily Basis ◽

Detection Methods ◽

Feed Forward Neural Network ◽

Training Scheme ◽

Training Schemes ◽

And Training

Introduction While surface-electromyography (sEMG) has been widely used in limb motion detection for the control of exoskeleton, there is an increasing interest to use forcemyography (FMG) method to detect motion. In this paper, we review the applications of two types of motion detection methods. Their performances were experimentally compared in day-to-day classification of forearm motions. The objective is to select a detection method suitable for motion assistance on a daily basis. Methods Comparisons of motion detection with FMG and sEMG were carried out considering classification accuracy (CA), repeatability and training scheme. For both methods, classification of motions was achieved through feed-forward neural network. Repeatability was evaluated on the basis of change in CA between days and also training schemes. Results The experiments shows that day-to-day CA with FMG can reach 84.9%, compared with a CA of 77.8% with sEMG, when the classifiers were trained only on the first day. Moreover, the CA with FMG can reach to 86.5%, comparable to CA of 84.1% with sEMG, if classifiers were trained daily. Conclusions Results suggest that data recorded from FMG is more repeatable in day-to-day testing and therefore FMG-based methods can be more useful than sEMG-based methods for motion detection in applications where exoskeletons are used as needed on a daily basis.

Download Full-text

DSP implmentation of a low complexity motion detection algorithm

Conference Record of the Thirty-Ninth Asilomar Conference onSignals, Systems and Computers, 2005. ◽

10.1109/acssc.2005.1599984 ◽

2006 ◽

Cited By ~ 2

Author(s):

P. Bassignana ◽

M. Martina ◽

G. Masera ◽

A. Molino ◽

F. Vacca

Keyword(s):

Motion Detection ◽

Low Complexity ◽

Detection Algorithm

Download Full-text

The complexity of DNA sequences. Different approaches and definitions

Математическая биология и биоинформатика ◽

10.17537/2020.15.313 ◽

2020 ◽

Vol 15 (2) ◽

pp. 313-337

Author(s):

Vladimir D. Gusev ◽

Liubov A. Miroshnichenko

Keyword(s):

Dna Sequences ◽

Quantitative Characteristic ◽

Low Complexity ◽

Structural Features ◽

General Definition ◽

Symbolic Sequence ◽

Different Types ◽

Definition Of

An important quantitative characteristic of symbolic sequence (texts, strings) is complexity, which reflects at the intuitive level the degree of their "non-randomness". A.N. Kolmogorov formulated the most general definition of complexity. He proposed measuring the complexity of an object (symbolic sequence) by the length of the shortest descriptions by which this object can be uniquely reconstructed. Since there is no program guaranteed to search for the shortest description, in practice, various algorithmic approximations considered in this paper are used for this purpose. Along with definitions of complexity, suggesting the possibility of reconstruction a sequence from its "description", a number of measures are considered that do not imply such restoration. They are based on the calculation of some quantitative characteristics. Of interest is not only a quantitative assessment of complexity, but also the identification and classification of structural regularities that determine its specific value. In one form or another, they are expressed in the demonstration of repetition in the broadest sense. The considered measures of complexity are conventionally divided into statistical ones that take into account the frequency of occurrence of symbols or short “words” in the text, “dictionary” ones that estimate the number of different “subwords” and “structural” ones based on the identification of long repeating fragments of text and the determination of relationships between them. Most of the methods are designed for sequences of an arbitrary linguistic nature. The special attention paid to DNA sequences, reflected in the title of the article, is due to the importance of the object, manifestations of repetition of different types, and numerous examples of using the concept of complexity in solving problems of classification and evolution of various biological objects. Local structural features found in the sliding window mode in DNA sequences are of considerable interest, since zones of low complexity in the genomes of various organisms are often associated with the regulation of basic genetic processes.

Download Full-text