Interest in utilizing neural networks a variety of scientific and academic studies industrial applications is increasing. In addition to the growing interest networks, there also rising video classification. Object detection from an image used as tool for various basis Identifying objects videos more difficult than single images, information has time continuity constraint. Common such ConvLSTM (Convolutional Long Short-Term Memory) 3DCNN (3D Convolutional Neural Network), well many others,...

In this paper, the performance of proposed Convolutional Neural Network (CNN) with three well-known image recognition methods such as Principal Component Analysis (PCA), Local Binary Patterns Histograms (LBPH) and K-Nearest Neighbour (KNN) is tested.In our experiments, overall accuracy PCA, LBPH, KNN CNN demonstrated.All experiments were implemented on ORL database obtained experimental results shown evaluated.This face consists 400 different subjects (40 classes/ 10 images for each...

Currently, three-dimensional convolutional neural networks (3DCNNs) are a popular approach in the field of human activity recognition. However, due to variety methods used for recognition, we propose new deep-learning model this paper. The main objective our work is optimize traditional 3DCNN and that combines with Convolutional Long Short-Term Memory (ConvLSTM) layers. Our experimental results, which were obtained using LoDVP Abnormal Activities dataset, UCF50 MOD20 demonstrate superiority...

In this paper, the Convolutional Neural Network (CNN) for classification of input animal images is proposed.This method compared with well-known image recognition methods such as Principal Component Analysis (PCA), Linear Discriminant (LDA), Local Binary Patterns Histograms (LBPH) and Support Vector Machine (SVM).The main goal to compare overall accuracy PCA, LDA, LBPH SVM proposed CNN method.For experiments, database wild animals created.This consists 500 different subjects (5 classes / 100...

In this paper a comparison between two popular feature extraction methods is presented. Scale-invariant transform (or SIFT) the first method. The Speeded up robust features SURF) presented as second. These are tested on set of depth maps. Ten defined gestures left hand in these Microsoft Kinect camera used for capturing images [1]. Support vector machine SVM) classification results accuracy SVM prediction selected images.

Recognizing various abnormal human activities from video is very challenging. This problem also greatly influenced by the lack of datasets containing activities. The available contain activities, but only a few them non-standard behavior such as theft, harassment, etc. There are KTH that focus on sudden behavioral changes, well changes in interpersonal interactions. UCF-crime dataset contains categories fighting, abuse, explosions, robberies, However, this time consuming. events videos occur...

This paper discusses research in the area of texture image classification. More specifically, combination and colour features is researched. The principle objective to create a robust descriptor for extraction features. principles two well-known methods grey-level feature extraction, namely GLCM (grey-level co-occurrence matrix) Gabor filters, are used experiments. For classification, support vector machine used. In first approach, applied separate channels image. experimental results show...

Bedsores are one of the severe problems which could affect a long-term lying subject in hospitals or hospice. To prevent bedsores, we present smart Internet Things (IoT) system for detecting position person using novel textile pressure sensors. build such system, it is necessary to use different technologies and techniques. We used sixty-four our sensors based on electrically conductive yarn Velostat collect information about distribution person. Using Message Queuing Telemetry Transport...

This study explores the extraction of remote Photoplethysmography (rPPG) signals from images using various neural network architectures, addressing challenge accurate signal estimation in biomedical contexts. The objective is to evaluate effectiveness different models capturing rPPG dataset snapshots. Two training strategies were investigated: pre-training with only fully connected layer being fine-tuned and entire scratch. analysis reveals that trained scratch consistently outperform their...

In this article, we are combining an advanced implementation of the popular ICP algorithm using transformation 3D invariant properties based on scale-invariant feature transform to register free-form closed surfaces (3D model human skull). Unlike point and surface registers, our method better captures bulk nature data such as bone thickness. The proposed is divided into three main steps: function extraction, comparison Euclidean metric distance gross alignment enhancement. input system...

In this paper, the performances of image recognition methods such as Principal Component Analysis (PCA), Linear Discriminant (LDA) and Local Binary Patterns Histograms (LBPH) are tested compared for input animal images. The main idea paper is to present an independent, comparative study some benefits drawbacks these most popular methods. Two sets experiments conducted relative performance evaluations. first part our experiments, accuracy PCA, LDA LBPH demonstrated. overall time execution...

In this paper, a novel method for object recognition based on hybrid local descriptors is presented. This utilizes combination of few approaches (SIFT - Scale-invariant feature transform, SURF Speeded Up Robust Features) and consists second parts. The applicability the presented methods are demonstrated images from dataset. Dataset classes represent big animals situated in Slovak country, namely wolf, fox, brown bear, deer wild boar. may be also used other areas image classification...

In the image analysis, segmentation is operation that divides into set of different segments. The paper deals about common color techniques and methods. advantages disadvantage each one will be described in this paper. At end paper, evaluation criterion introduced applied on algorithms results. Five most used algorithms, namely, Efficient graph based, K-means, Mean shift, Expectation maximization hybrid method are compared by designed criterion.

In this paper, a novel system for automatic detection and classification of animal is presented. System called ASFAR (Automatic For Animal Recognition) based on distributed so-called `watching device' in designated area main computing unit (MCU) acting as server manager. Watching devices are situated wild nature their task to detect then send data MCU evaluation. The whole determine migration corridors animals area. To create object representation, visual descriptors were chosen Support...

This article discusses the creation of adaptive 3D video streaming that adjusts to various internet speeds and devices. It employs tools such as HLS open-source software like FFmpeg, Nginx, Video.js. Emphasizing critical role in optimizing performance, particularly content, it draws inspiration from applications integrating medical training for neurosurgery. The details a specific project utilizing these tools, with primary focus on ensuring smooth playback. delves into process developing...

This paper presents a proposed methodology for face recognition based on an information theory approach to coding and decoding images. In this paper, we propose 2D-3D face-matching method principal component analysis (PCA) algorithm using canonical correlation (CCA) learn the mapping between 2D image 3D data. makes it possible match with enrolled Our fusion is PCA method, which applied extract base features. feature-level requires extraction of different features from source data before are...

In this paper, we present an assessment framework that can be used to score segments of physical and digital infrastructure based on their features readiness expedite the deployment Connected Automated Vehicles (CAVs). We discuss equipment methodology applied for collection analysis required data in automated way. Moreover, demonstrate how proposed using collected a public transport route city Zilina, Slovakia. use two types assessment-connectivity positioning assess connectivity...

In this paper, a 3D reconstruction algorithm using CT slices of human pelvis is presented. We propose the method for image that based on combination SURF (Speeded-Up Robust Features) descriptor and SSD (Sum Squared Differences) matching segmentation with aim to obtain accurate model pelvis. Firstly, we apply filtering noise removing smoothing. Next, filtered split into segments Mean- Shift algorithm. Secondly, edges Canny edge detector are extracted. Then, each segment look at associated...

This paper proposes a 3D surface registration algorithm based on the iterated closest point (ICP). The proposed uses Scale-Invariant Feature Transform (SIFT) functions for initial alignment in combination with K-Nearst Neighbor (KNN) function comparison and Iterative Closest Point (ICP) weighted performing accurate registration. First, area properties are used corresponding cloud areas. Second, files associated regions classified to calculate transformation matrix. Based this combination,...

This paper provides an example of the face recognition using SIFT-PCA method and impact Graph Based segmentation algorithm on rate. Principle component analysis (PCA) is a multivariate technique that analyzes data in which observation are described by several inter-correlated dependent variables. The goal to extract important information from data, represent it as set new orthogonal variables called principal components. presents proposed methodology for based preprocessing images SIFT...

This paper provides a new feature extraction method for object recognition using PCA-KNN algorithm with SIFT descriptor. The proposed is divided into three steps. first step based on from the input images (Scale Invariant Feature Transform) Each of features represented one or more descriptors. In medical systems used as patterns are also by vectors. second eigen values and vectors extracted each image. We apply PCA after we reduce number algorithm. goal to extract important information set...

In this paper, the comparison between deep learning methods and feature extraction algorithms is presented. The principle of Grey-Level Co-occurrence Matrix (GLCM) its modifications are used for our research. main idea was to design a method description combined features textures. texture classification process carried out with robust support vector machine classifier (SVM). We compare these proposed Convolutional Neural Networks (CNN). This network contains 25 layers. Finally, all...

This article is focused on the automatic classification of passing vehicles through an experimental platform using optical sensor arrays. The amount data generated from various systems growing proportionally every year. Therefore, it necessary to look for more progressive solutions these problems. Methods implementing artificial intelligence are becoming a new trend in this area. At first, with two separate groups fiber Bragg grating arrays (horizontally and vertically oriented) installed...