A REVIEW ON FACIAL EMOTION RECOGNITION THAT USES MACHINE LEARNING ALGORITHMS

For a computer, identification of human emotion from a still image of the human face is a complex, challenging, and heavily calculative task. Classification of human emotion is done by using a different combination of convolutional neural networks (CNN) that task is known as Facial Emotion Recognition (FER). CNN model is achieved by training and testing on lots of same categorical images from the dataset using different hyperparameter tuning. The main contribution of this work is to look for various CNN architectures, hyperparameter tuning and compare the performance of those CNN models based on accuracy and loss while training and testing on Facial Emotion Recognition. This study shall help to provide a guide for the selection of an appropriate CNN model and tuning parameter according to the needs of the applicant.

INVESTIGATION AND DEVELOPMENT OF “UNIVERSAL IMAGE DICTIONARY” FOR CREATION OF MAN-MACHINE INTERFACE

The project urgency is evidenced by the measurement of subject-matter popularity on the Internet – a query to the Net on the topic of the subject gives 18900 results, and a novelty is defined by specification, for instance, the “Universal Image Dictionary” query gives “No results found for” and 1 result M24.RU – 10 unknown: monsters, stairs, funnels and keys… www.m24.ru/articles/112668/ … The project offered supposes the creation of the bank of different images already existing and having a wide distribution and which can be a means of international communication for people having no any other channel for information exchange. Such images could include common gestures, traffic signs, signs in transport, in the streets, in public accommodations and in state offices. They will include both single images, and their combinations forming a single conceptual complex (rules of table etiquette, on transport, and at the stadium etc.). It should be emphasized clearly that the dictionary offered is intended for interpersonal communication. A computer identification of images cannot be in this case a basic purpose for project realization. The orientation to interpersonal communication gives us a possibility to choose images for the dictionary and, what is not less significant, to set problems actually solvable at every stage of its creation.

Clock Skew Based Computer Identification on Different Types of Area Networks

The increase of client devices along with the growth of internet access currently affects to security threats at the user's identity. Identifiers that commonly used today, such as SSID, IP address, MAC address, cookies, and session IDs have a weakness, which is easy to duplicate. Computer identification based on clock skew is an identification method that is not easily duplicated because it is based on the hardware characteristics of the device. Clock skew is the deviation of the clock to the true time which causes each clock to run at a slightly different speed. This study aims to determine the effect of network types to the clock skew stability as a reliable device identification method. This research was conducted on five client computers which running windows and linux operating systems. The measurement was conducted based on three different types of area networks, i.e., LAN, MAN, and WAN. The skew estimation was done using two linear methods i.e., linear programming and linear regression. The measurement results show that the most stable clock skew is found on the LAN measurement because it meets the threshold tolerance limit i.e., ±1 ppm. Skew estimation using linear programming method has better accuracy than linear regression method.

Classification of traffic signal system anomalies for environment tests of autonomous vehicles

In the future there will be a lot of changes and development concerning autonomous transport that will affect all participants of transport. There are still difficulties in organizing transport, but with the introduction of autonomous vehicles more challenges can be expected. Recognizing and tracking horizontal and vertical signs can cause a difficulties for drivers and, later, for autonomous systems. Environmental conditions, deformity and quality affect the perception of signals. The correct recognition results in safe travelling for everyone on the roads. Traffic signs are designed for people that is why the recognition process is harder for the machines. However, nowadays some developers try to create a traffic sign that autonomous vehicles can use. Computer identification needs further development, as it is necessary to consider cases where traffic signs are deformed or not properly placed. In the following investigation, the advantages and disadvantages of the different perception methods and their possibilities were gathered. A methodology for the classification of horizontal and vertical traffic signs anomalies that may help in designing better testing and validation environments for traffic sign recognition systems in the future was also proposed.

Dichroism Infrared-reflective Thermochromic Ink of Mobile Terminals’ Recognition

The decoration of thermochromic ink has applications ranging from engineering devices to liquid sensors. However, when it comes to accurate computer identification the characteristic as dichroism is not fully used. Thus, we have recently proposed a manufacture method of dichroism infrared-reflective thermochromic ink. (DITI). By rubbing to heat DITI coated layer, this method enables us to finely and separately create dichroism manuscript after the handwriting is complete. Moreover, the microcapsules have cheerful compatibility. We have decorated conventional microcapsules with bola-type copolymer which allow us to fill them with Infrared (IR) reflective perssad. We consider this method promising to distinguish dichroism printing or writing on a tanglesome surface much quicker, for the IR reflective low-colour-density layer will be easily locked by an IR camera or a mobile phone camera to adjust specific characters of interest. We anticipate that this DITI ink will find use in markable temperature measurement, fluid machinery, anti-counterfeiting techniques and daily printing or writing.

EXPRESS PREDICTION OF EXTERNAL DISTINCTIVE FEATURES OF PERSON USING THE PROGRAM OF DERMATOGLYPHICS FOR PREDICTION

The aim of our study was to investigate the current state of computer identification applications, such as artificial neural networks. The material of our study were antroposcopic and anthropometric parameters obtained from 180 male and females aged 18–55 years living in the Ivano-Frankivsk region and belonging to Boiko, Lemko or Hutsul ethno-territorial group. Prints of comb pattern of the toes obtained by scanning with Futronic's FS80 USB2.0 Fingerprint Scanner using the program ftrScanApiEx.exe. followed by the transfer of data to a personal computer. For statistical processing of the obtained data we use STATISTICA 12 from the company StatSoft. Construction of neural networks was carried out using Neural Networks. As a result of our research there was carried out the prediction of anthropometric and antroposcopic parameters (ethno-territorial and gender belonging, etc.) through the use of dermatoglyphic parameters of the hands and feet in 180 people living in the Ivano-Frankivsk region. The proposed method allowed to obtain the results with a forecasts probability 73–90 %. The use of above algorithm of actions allowed a 50 % increase of quality of identification of unknown person for using dermatoglyphic method and 67 % facilitatation of the process of identification (of quantitative and qualitative calculations, determining correlations between parameters) in comparison with previously known manner. Therefore, our proposed method can be used as an express diagnostics of common phenotypic traits of the person (ethno-territorial affiliation, gender, etc.) at admission of mass victims (natural disasters, acts of terrorism, armed conflicts, man-made disasters, etc.), it doesn’t not require a long time for conducting, specially trained staff and is inexpensive.Conclusions: The possibility of predicting external-recognizing features of a person such as etno-racial belonging, sex, anthropometric and antroposcopic parameters will allow widely use dermatoglyphic method at the level with other methods in conducting forensic identification of impersonal, fragmented and putrefactive modified corpses.

Geometric and Textural Cues for Automatic Kinship Verification

Automatic Kinship verification aims at recognizing the degree of kinship of two individuals from their facial images and it has possible applications in image retrieval and annotation, forensics and historical studies. This is a recent and challenging problem, which must deal with different degrees of kinship and variations in age and gender. Our work explores the computer identification of parent–child pairs using a combination of (i) features of different natures, based on geometric and textural data, (ii) feature selection and (iii) state-of-the-art classifiers. Experiments show that the proposed approach provides a valuable solution to the kinship verification problem, as suggested by its comparison with different methods on the same data and the same experimental protocols. We further show the good generalization capabilities of our method in several cross-database experiments.