International Journal of Image, Graphics and Signal Processing @ijigsp
Journal articles - International Journal of Image, Graphics and Signal Processing
All articles: 1220
Scientific article
Crankshaft is the most important and hard manufacture part in engine and diesel, its deformation contains multiple arcs, and its straightening technology is a complex NP-hard determination problem. The crankshaft straightening process was analyzed, and some effects from multiple arcs crankshaft straightening was investigated. The mathematics model of multiple arcs straightening technology determination system was built based on graph theory principle. And the Influence of Bauschinger effect was considered, the online straightening technology determination calculation algorithm based on weighted evaluation function was proposed. The experiments about straightening crankshaft were performed in YH40-160 straightening press to evaluate two different weighted set C1={2.5,1.5,1,1.2,3.5} and C2={1,1,1,1,1}, and the results expressed that the online straightening technology determination calculation algorithm was advantage to crankshaft straightening process, and C1 had better influence to qualities and efficiencies than C2.
Free
Crop Type Classification Based on Clonal Selection Algorithm for High Resolution Satellite Image
Scientific article
This paper presents a hierarchical clustering algorithm for crop type classification problem using multi-spectral satellite image. In unsupervised techniques, the automatic generation of clusters and its centers is not exploited to their full potential. Hence, a hierarchical clustering algorithm is proposed which uses splitting and merging techniques. Initially, the splitting method is used to search for the best possible number of clusters and its centers using non-parametric technique i.e., clonal selection method. Using these clusters, a merging method is used to group the data points based on a parametric method (K-means algorithm). The performance of the proposed hierarchical clustering algorithm is compared with two unsupervised algorithms (K-means and Self-Organizing Map) that are available in the literature. A performance comparison of the proposed algorithm with the conventional algorithms is presented. From the results obtained, we conclude that the proposed hierarchical clustering algorithm is more accurate.
Free
Crowd Behaviour Analysis for Enhanced Event Safety and Management
Scientific article
Addressing crowd control and safety at large-scale events is the central focus of this study. The proposed methodology is tested on ShanghaiTechA, ShanghaiTechB and UCF CC50 datasets. Apart from VGG-16 referred as the baseline model, the study utilizes a Convolutional Neural Network (CNN) model like VGG with dilatable layers and Atrous Spatial Pyramid Pooling (ASPP) layers on these datasets to identify every individual in the crowd by their heads. Furthermore, optical flow analysis identifies fast-moving pixels, facilitating the detection of rapid movements within the crowd. YOLO tracking is additionally employed to monitor the direction of object movement within the crowd. By integrating these methodologies, the study aims to enhance overall safety and security of individuals in the crowd. VGG with dilatable layers gives the least Mean Absolute Error for ShanghaiTechA and ShanghaiTechB datasets. The ASPP approach demonstrates approximately 15% higher accuracy on average compared to the baseline model for the ShanghaiTechA and UCF CC 50 datasets.
Free
Scientific article
In this paper we propose a method for automatic detection of crowd escape behaviour. Motion features are extracted by optical flow using Lucas-Kanade derivative of Gaussian method (LKDoG) followed by robust probabilistic weighted feature pooling operation. Probabilistic feature polling chooses the most descriptive features in the sub-block and summarizes the joint representation of the selected features by Probabilistic Weighted Optical Flow Magnitude Histogram (PWOFMH) and Probabilistic Weighted Optical Flow Direction Histogram (PWOFDH). One class Extreme Learning Machine (OC-ELM) is used to train and test our proposed algorithm. The accuracy of our proposed method is evaluated on UMN, PETS 2009 and AVANUE datasets and correlations with the best in class techniques approves the upsides of our proposed method.
Free
Curvelet transform for efficient static texture classification and image fusion
Scientific article
Wavelet Transform (WT) has widely been used in signal processing. WT breaks a signal into its wavelets that are scaled and shifted versions of given signal. Thus wavelets are able represent graphical objects. The irregular shape and compact support of wavelets made them ideal for analyzing non-stationary signals. They are useful in analysis in both temporal and frequency domains. In contract, the Fourier transform provides information in frequency domain lacking in information in time domain. Thus wavelets became popular for signal processing and image processing applications. Nevertheless, wavelets suffer from a drawback as they cannot effectively represent images at different angles and different scales. To overcome this problem, of late, Curvelet Transform (CT) came into existence. CT is nothing but the higher dimensional generalization of WT which can effectively represent images at different angles and different scales. In this paper we proposed a CT method that is used to represent textures and classify them. The methodology used in this paper has an underlying approach that exploits statistical features of curvelets that resulted in curvelet decomposition. We built a prototype application using MATLAB to demonstrate proof of the concept.
Free
Curvilinear Tracing Approach for Extracting Kannada Word Sign Symbol from Sign Video
Scientific article
Gesture based communications are utilized as a primary method of correspondence, however, the differing qualities in the sign image portrayal limit its use to district bound. There is a tremendous assorted quality in the sign image portrayal from one nation to another, one state to another. In India, there is distinctive gesture-based communication watched for each state locale. It is henceforth exceptionally troublesome for one area individual to convey to other utilizing a signature image. This paper proposes a curvilinear tracing approach for the shape portrayal of Kannada communication via gestures acknowledgment. To build up this approach, a dataset is consequently made with all Swaragalu, Vyanjanagalu, Materials and Numbers in Kannada dialect. The arrangement of the dataset is framed by characterizing a vocabulary dataset for various sign images utilized as a part of regular interfacing. In the portrayal of gesture-based communication for acknowledgment, edge elements of hand areas are thought to be an ideal element portrayal of communication through signing. In the preparing of gesture-based communication, the agent includes assumes a critical part in arrangement execution. For the developed approach of sign language detection, where a single significant transformation is carried out, a word level detection is then performed. To represent the processing efficiency, a set of cue symbols is used for formulating a word. This word symbols are then processed to evaluate the performance for sign language detection. Word processing is carried out as a recursive process of a single cue symbol representation, where each frame data are processed for a curvilinear shape feature. The frame data are extracted based on the frame reading rate and multiple frames are processed in successive format to extract the region of interest. A system outline to process the video data and to give an optimal frame processing for sign recognition a word level process is performed.
Free
