Статьи журнала - International Journal of Image, Graphics and Signal Processing
Все статьи: 1110
Статья научная
In medical diagnosis, Artificial Intelligence (AI) has offered significant revolution, especially for cancers. Acute Myeloid Leukemia (AML) is a deadly blood cancer caused by the rapid growth of abnormal White Blood Cells (WBCs) in humans. Although AML classification is a popular area of research, existing detection methods utilize manual examination of microscopic blood samples, which includes high complexity and tedious. Therefore, this work presented a computerized deep learning model-based AML classification from peripheral blood stain images, which helps in earlier AML diagnosis. The processing steps followed in AML classification are Image Pre-processing, Localization of RoI (Region of Interest), Fusion-based Feature Extraction and Classification. First, the input image is pre-processed, which includes noise filtering, image resizing, and colour conversion. The noise in the image is filtered using normalized Gaussian filtering (NGF). Next, the image is resized into a standard size, and the RGB image is converted into CMYK colour space. Then, the RoI is identified using the Image Moment Localization (IML) technique. Next, the valuable multi-level dense features are extracted using DenseSqueeze Network, and multi-scale features are extracted using Dilated Convolution Spatial Pyramid Pooling (Dilated CSPP). Both these extracted features are fused using the element-wise summation. Finally, the Softmax classifier is used in the last layer to classify the classes of AML and the loss in the network is optimized using the Improved Artificial Fish Swarm (Improved AFS) algorithm. The proposed work results in 99% of accuracy, 98.5% of precision and 98.9% of F-score by using the AML-Cytomorphology LMU dataset.
Бесплатно
Contact distribution function based clustering technique with self-organizing maps
Статья научная
Currently clustering techniques play a vital role in object recognition process. The clustering techniques are found to be efficient with neural networks. So, the present paper proposed a novel method for clustering the input objects with Self-Organizing Map (SOM). The proposed method considers the input object as a random closed set. The random set can be efficiently described with various features viz., volume fractions, co-variance and contact distributions etc. In the proposed method, the input object is described efficiently with spherical contact distribution. The proposed method is experimented with the leaf data set with 795 images. The performance of the proposed method is evaluated with various topologies of SOM and is measured with four measures viz., FNR, FPR, TPR and TNR. The results indicate the efficiency of the proposed method.
Бесплатно
Content based image retrieval using multi motif co-occurrence matrix
Статья научная
In this paper, two extended versions of motif co-occurrence matrices (MCM) are derived and concatenated for efficient content-based image retrieval (CBIR). This paper divides the image into 2 x 2 grids. Each 2 x 2 grid is replaced with two different Peano scan motif (PSM) indexes, one is initiated from top left most pixel and the other is initiated from bottom right most pixel. This transforms the entire image into two different images and co-occurrence matrices are derived on these two transformed images: the first one is named as “motif co-occurrence matrix initiated from top left most pixel (MCMTL)” and second one is named as “motif co-occurrence matrix initiated from bottom right most pixel (MCMBR)”. The proposed method concatenates the feature vectors of MCMTL and MCMBR and derives multi motif co-occurrence matrix (MMCM) features. This paper carried out investigation on image databases i.e. Corel-1k, Corel-10k, MIT-VisTex, Brodtaz, and CMU-PIE and the results are compared with other well-known CBIR methods. The results indicate the efficacy of the proposed MMCM than the other methods and especially on MCM [19] method.
Бесплатно
Content-based Search for Image Retrieval
Статья научная
In this paper, a content-based image retrieval approach is presented for effective searching. The proposed approach uses two or more types of query for accessing images, textual annotation associated with images and visual appearance, such as colour, texture and positional features of objects in sample images. One can first place a keyword-based query, and then the desired images are retrieved by visual content-based query. The proposed retrieval approach shows clear improvements over competing approaches in terms of retrieval accuracy and visual inspection using Corel gallery and WWW images.
Бесплатно
Contour Based Retrieval for Plant Species
Статья научная
Recognizing a plant in any huge vegetation is a tedious work for us. We recognize a plant on the basis of its size, leaves, flowers, fruits, etc. Leaf is a part of the plant which can be found on plants almost in all seasons and most of the time we have to recognize plants on the basis of its leaf. But when dealing with leaf of plant, it is important to consider the finer details of the contour representing the shape of the leaf. We are trying to build a system which has a database of leaves of different plants and given a leaf, we find out the plant to which it may belong. In this paper, we present the results of tangential angle approach used for retrieval. A database of around one thousand leaves of different plants has been created. Each leaf image is preprocessed to extract its boundary. Then tangential angle approach is applied which captures the angular details of the boundary of shape. We have done the testing for around 1000 leaves and on the basis of that recall, precision and error rate have been calculated to measure the effectiveness of the proposed method.
Бесплатно
Contrast Enhancement of Images through Skewness and Mode Based Bi-Histogram Equalization
Статья научная
In this paper, skewness and mode-based histogram equalization algorithm have been proposed for contrast enhancement of digital images. The present method gives a novel idea for histogram clipping and histogram bifurcation. The prior is done with the skewness value and the latter is done with help of mode values of the intensity level random data set. The pixel intensity levels are random and thus a stochastic approach has been used and found to yield improved figure of merits. The image histogram has been clipped with the help of a pre-assigned threshold value computed from skewness value to restrict the rate of over enhancement. The clipped histogram is subdivided into two parts, using the histogram subdivision limit which is calculated on the basis of the mode value of the image. Histogram of individual sub-image is equalized independently and then integrated to form the final enhanced image. The simulation results have shown that the proposed skewness and mode based bi-histogram equalization algorithm enhances the contrast of the image in a better manner compared with the other histogram equalization methods in terms of FSIM, PSIM, SFF, VSI, HaarPSI, and GMSD.
Бесплатно
Contribution to the Fusion of Biometric Modalities by the Choquet Integral
Статья научная
In multimodal biometrics, modalities can be robust against the authentication of certain people and weak for others. The conventional fusion techniques such as the Product, Mean, AND, OR and the Majority Voting do not take into account this kind of behaviour. In this paper, we propose a new approach to fusion procedures in the context of biometric authentication. The proposed method is based on the exploration of the Choquet integral that takes into account the interactions between the terms and people through fuzzy measures. The fuzzy measures, the ones we have proposed, are based on the number of confusion, the entropy and the uncertainty function. The results have been validated in two databases: the first one is virtual, which is based on synthetic scores and the second one on the biometric modalities which are: face, off-line handwriting and off-line signature. The achieved results demonstrate the robustness of our approaches and their adaptability.
Бесплатно
Convolution Based Technique for Indic Script Identification from Handwritten Document Images
Статья научная
Determination of script type of document image is a complex real life problem for a multi-script country like India, where 23 official languages (including English) are present and 13 different scripts are used to write them. Including English and Roman those count become 23 and 13 respectively. The problem becomes more challenging when handwritten documents are considered. In this paper an approach for identifying the script type of handwritten document images written by any one of the Bangla, Devnagari, Roman and Urdu script is proposed. Two convolution based techniques, namely Gabor filter and Morphological reconstruction are combined and a feature vector of 20 dimensions is constructed. Due to unavailability of a standard data set, a corpus of 157 document images with an almost equal ratio of four types of script is prepared. During classification the dataset is divided into 2:1 ratio. An average identification accuracy rate of 94.4% is obtained on the test set. The average Bi-script and Tri-script identification accuracy rate was found to be 98.2% and 97.5% respectively. Statistical performance analysis is done using different well known classifiers.
Бесплатно
Статья научная
Fires spread quickly and are extremely difficult to contain, and cause a great deal of damage to people and property. Current domestic systems for detecting outbreaks of fire, such as smoke detectors, are prone to reliability issues and will benefit greatly from having a secondary system in place to confirm the presence of a fire in the premises. In this paper, we have proposed a novel image pre-processing algorithm known as the Selective Amplification. This technique enhances images that are to be used in Convolutional Neural Networks, which are then trained on pre-processed images to detect fires with high accuracy. The efficacy of the proposed technique is verified by training two identical Convolutional Neural Network models on the same dataset of images. We train the proposed model on a version of the dataset that uses Selective Amplification for data pre-processing. The proposed model then demonstrates an improvement in the accuracy of the detection of fire in real-time over by 12.85%, compared to an identical model trained on the dataset without any pre-processing performed beforehand.
Бесплатно
Convolutional Neural Network based Handwritten Bengali and Bengali-English Mixed Numeral Recognition
Статья научная
Recognition of handwritten numerals has gained much interest in recent years due to its various potential applications. Bengali is the fifth ranked among the spoken languages of the world. However, due to inherent difficulties of Bengali numeral recognition, a very few study on handwritten Bengali numeral recognition is found with respect to other major languages. The existing Bengali numeral recognition methods used distinct feature extraction techniques and various classification tools. Recently, convolutional neural network (CNN) is found efficient for image classification with its distinct features. In this paper, we have investigated a CNN based Bengali handwritten numeral recognition scheme. Since English numerals are frequently used with Bengali numerals, handwritten Bengali-English mixed numerals are also investigated in this study. The proposed scheme uses moderate pre-processing technique to generate patterns from images of handwritten numerals and then employs CNN to classify individual numerals. It does not employ any feature extraction method like other related works. The proposed method showed satisfactory recognition accuracy on the benchmark data set and outperformed other prominent existing methods for both Bengali and Bengali-English mixed cases.
Бесплатно
Copy Move Forgery Detection in Contrast Variant Environment using Binary DCT Vectors
Статья научная
Copy move forgery detection is a very popular research area and a lot of methods have been suggested by researchers. However, every method has its own merits and weaknesses and hence, new techniques are being continuously devised and analyzed. There are many post processing operations used by the manipulators to obstruct the forgery detection techniques. One such operation is changing the contrast of the whole image or copy moved regions, which many existing methods fail to address. A novel method using binary discrete cosine transform vectors is proposed to detect copy move forgery in the presence of contrast changes. The image is divided into overlapping blocks and DCT coefficients are calculated for these blocks. Feature vectors are created from these blocks using signs of the DCT coefficients. Coefficient of correlation is used to match resulting binary vectors. The experiments show that the proposed method is able to detect copy move forgery in presence of contrast changes. The proposed method is also invariant to other post processing operations like Gaussian noise, JPEG compression and little rotation and scaling.
Бесплатно
Статья научная
Due to the availability of low-cost electronic devices and advanced image editing tools, changing the semantic meaning of a particular image has become straightforward by employing various image manipulation techniques like image copy-move, image splicing and removal operations. The tampered images with this sophisticated software are rich in visualization, making the modifications invisible to the naked eye. Detecting these image alterations is laborious, time-consuming, and often yields inappropriate results. The current techniques use conventional square, slide regular, and artifacts procedures to identify image deviations to combat image forgery practices. Still, these techniques exhibit problems related to generalization, training and testing, and model complexity. So, in this paper, a novel image forgery detection and localization framework is implemented using stationary wavelet transform (SWT), and a Hybrid Dilated Adaptive VGG16 model with optimization is introduced to classify forgery images and localize the forgery regions present in an image. Initially, the proposed framework processes the input image with SWT to decompose an image into different subband and further divide it into patches. After that, the hybrid dilated adaptive VGG16 Network (HDA-VGG16Net) is built to extract the deep image features from the patches. Later, the Hybridized Tuna Swarm with Bald Eagle Search Optimization (HTS-BESO) technique is applied to optimize the VGG16 parameters. Finally, feature matching is formed using multi-similarity searching to recognize whether the input image is forged or original by locating forgery regions. The evaluation results are compared with existing forgery detection approaches to ensure the efficiency of the developed model by considering multiple performance measures.
Бесплатно
Copy-Move Image Forgery Detection a Review
Статья научная
Due to the availability of various image processing tools forgery over an image can be performed very easily but very difficult to identify. In copy-move forgery, a segment is copied from the original image and pasted at some other location on the same image to hide significant objects of image or to bring additional information which is originally not present in image. Nowadays, this forgery technique is drawing researcher's attention. Till now many solutions are presented by researchers to detect such type of forgery in images. Several post-processing operations like rotation, alteration in intensity, noise addition, filtering and blurring can be applied over copy-moved segment which makes detection of forgery very difficult. Copy-move forgery detection is mainly based on finding similarity present in an image and establish a relationship between genuine image parts and pasted portion of the image. This paper is centralized towards providing survey to forgery detection techniques based on different block-based methods. In block-based methods image is divided in blocks of fixed dimension and further features are extracted corresponding to each block of image. Forged blocks are identified utilizing the similarity present between feature vectors.
Бесплатно
Correcting Multi-focus Images via Simple Standard Deviation for Image Fusion
Статья научная
Image fusion is one of the recent trends in image registration which is an essential field of image processing. The basic principle of this paper is to fuse multi-focus images using simple statistical standard deviation. Firstly, the simple standard deviation for the kk window inside each of the multi-focus images was computed. The contribution in this paper came from the idea that the focused part inside an image had high details rather than the unfocused part. Hence, the dispersion between pixels inside the focused part is higher than the dispersion inside the unfocused part. Secondly, a simple comparison between the standard deviation for each kk window in the multi-focus images could be computed. The highest standard deviation between all the computed standard deviations for the multi-focus images could be treated as the optimal that is to be placed in the fused image. The experimental visual results show that the proposed method produces very satisfactory results in spite of its simplicity.
Бесплатно
Cosine Modulated Non-Uniform Filter Bank with Improved Computational Efficiency
Статья научная
In this correspondence, a simple design method for multirate near perfect reconstruction (NPR) integer decimated filter bank with non-uniform frequency spacing and linear phase property, that involves optimization of only single parameter, is proposed. It is derived from the uniform cosine modulated filter bank (CMFB) by merging some relevant band pass filters. The design procedure and the structure of the uniform CMFB are mostly preserved in the non-uniform implementation. The parent filter of the filter bank is formulated as an interpolated finite impulse response (IFIR) filter. The IFIR digital filters permit efficient hardware implementations due to less number of multiplier coefficients. Design examples show that the proposed approach provides good performance with less computational complexity at the cost of slight increase in system delay.
Бесплатно
