Kussay N. Mutter

Automatic Fingerprint Identification Using Gray Hopfield Neural Network Improved by Run-Length Encoding

This paper presents a new technique of fingerprint identification using gray Hopfield neural network (GHNN) improved by run-length encoding (RLE). Gabor filter has been used for image enhancement at the stage of enrollment and vector field algorithm for core detection as a reference point. Finding this point will enable to cover most of information around the core. GHNN deals with gray level images by learning on bitplanes that represent the fingerprint image layers. For large number of images GHNNs memory needs very large storage space to cover all learned fingerprint images. RLE is a very simple and useful solution for saving the capacity of the net memory by encoding the stored weights, in which the weights data will reduce according to the repeated one. Experiments carried out on fingerprint images show that the proposed technique is useful in a number of different samples of fingerprint images in terms of converged images in quality, encoding and decoding performance.

Gray Image Recognition Using Hopfield Neural Network With Multi-Bitplane and Multi-Connect Architecture

In this work, a method for applying Hopfield neural network (HNN) with gray images is presented. Hopfield networks are iterative auto-associative networks consisting of a single layer of fully connected processing elements thus categorizes as an associative memory. Associative memories provide one approach to the computer-engineering problem of storing and retrieving data which is based on content rather than storage address. HNN deals with the bipolar system (i.e. -1 and +1) for direct input data, however it is useful for binary images, but unuseful for gray-level or color images unless we suppose another way for input data of such images. To overcome this obstacle, one can suppose for 8-bit gray-level image that consists of 8-layers (bitplanes) of binaries can be represented as bipolar data. In this way it is possible to express each bitplane as single binary image for HNN. The experimental results showed the usefulness of using HNN in gray-level images recognition with good results. Furthermore, there are no limitations to the number of 8-bit gray level images that can be stored in the net memory with the same efficient results

Segmentation and estimation of brain tumor volume in computed tomography scan images using hidden Markov random field Expectation Maximization algorithm

Brain tumors have been created by abnormal and uncontrolled cell division inside the brain. A crucial and lengthy task is the segmentation of brain tumors, which can be gained manually with the help of Computed Tomography (CT). Treatment, diagnosis, signs and symptoms of the brain tumors mainly depend on the volume, shapes and location of the tumors. The accuracy and time of detecting brain tumor are vital contributions in the successful diagnosis and treatment of tumors. Therefore, the detection of brain tumor needs to be fast and accurate. Brain tumor segmentation and volume estimation have been considered a challenge mission in medical image processing. The main aim of this paper is that with the help of hidden Markov random field- Expectation Maximization (HMRF-EM) and threshold method, a novel approach of improving the segmentation of brain tumors from CT scan images is produced. The segmentation and volume estimation images are obtained by the study of 2D images. We calculate the volume of tumor using a new approach based on 2D images estimations and voxel space. In order to validate the proposed approach a comparison is carried out with a manual method using Mango software which, the noise or impurities are less than Mango software in measurement of tumor volume.

A Modified Hopfield Neural Network Algorithm (MHNNA) Using ALOS Image for Water Quality Mapping

Decreasing water pollution is a big problem in coastal waters. Coastal health of ecosystems can be affected by high concentrations of suspended sediment. In this work, a Modified Hopfield Neural Network Algorithm (MHNNA) was used with remote sensing imagery to classify the total suspended solids (TSS) concentrations in the waters of coastal Langkawi Island, Malaysia. The adopted remote sensing image is the Advanced Land Observation Satellite (ALOS) image acquired on 18 January 2010. Our modification allows the Hopfield neural network to convert and classify color satellite images. The samples were collected from the study area simultaneously with the acquiring of satellite imagery. The sample locations were determined using a handheld global positioning system (GPS). The TSS concentration measurements were conducted in a lab and used for validation (real data), classification, and accuracy assessments. Mapping was achieved by using the MHNNA to classify the concentrations according to their reflectance values in band 1, band 2, and band 3. The TSS map was color-coded for visual interpretation. The efficiency of the proposed algorithm was investigated by dividing the validation data into two groups. The first group was used as source samples for supervisor classification via the MHNNA. The second group was used to test the MHNNA efficiency. After mapping, the locations of the second group in the produced classes were detected. Next, the correlation coefficient (R) and root mean square error (RMSE) were calculated between the two groups, according to their corresponding locations in the classes. The MHNNA exhibited a higher R (0.977) and lower RMSE (2.887). In addition, we test the MHNNA with noise, where it proves its accuracy with noisy images over a range of noise levels. All results have been compared with a minimum distance classifier (Min-Dis). Therefore, TSS mapping of polluted water in the coastal Langkawi Island, Malaysia can be performed using the adopted MHNNA with remote sensing techniques (as based on ALOS images).

Modified Hopfield Neural Network Algorithm (MHNNA) for TSS mapping in Penang strait, Malaysia

The use of traditional ship sampling method of for environmental monitoring is time consuming, requires a high survey cost, and exert great efforts. In this study we classify one of the water pollutants which is the Total Suspended Solids (TSS) of polluted water in Penang strait, Malaysia by applying Modified Hopfield Neural Network Algorithm (MHNNA) on THEOS (Thailand Earth Observation System) image. The samples were collected from study area simultaneously with the airborne image acquisition. The samples locations were determined by using a handheld global positioning system (GPS), and the measurement of TSS concentrations was conducted in the lab as validation data (sea-truth data). By using algorithm (MHNNA) the concentrations of TSS have been classified according their varied values to produce the map. The map was colour-coded for visual interpretation. The investigation of efficiency of the proposed algorithm was based on dividing the validation data into two groups, the first group refers to standard samples for supervisor classification by the used algorithm. And the second group for test, where after classification we detect the second group data positions in the produced classes, then finding correlation coefficient (R) and root-mean-square-error (RMSE) between the first group data and the second group data according to their correspondence in the classes. The observations were high (R=0.899) with low (RMSE=17.687). This study indicates that TSS mapping of polluted water can be carried out using remote sensing technique by the application of MHNNA on THEOS satellite data over Penang strait, Malaysia.

Real time object detection using Hopfield neural network for Arabic printed letter recognition

In this work, a new technique of improving Hopfield model for object edge detection of Arabic letters recognition is proposed. In conventional methods, different trends for object segmentation are used to split cursive letters individually for recognition. The presented technique differentiates only letters with no maintain of background data. Each letter is a set of clustered small weights distributed according to its shape within the word. The average of Total Letter Weight is a special property for each form of the letters. Preliminary experimental tests show positive performance of the proposed system.

Hopfield neural network (hnn) improvement for color image recognition using multi-bitplane and multi-connect architecture

A new approach of using HNN with multi-connect architecture in color image recognition has been produced in this work. HNN consists of a single layer of fully connected processing elements, which is described as an associative memory. However, HNN is useless in dealing with data not in bipolar representation. As such, HNN failed to work directly with color images, unless, another way is produced in order to pave the way for expected right recognition. In RGB bands each represents different values of brightness, still it is possible to assume for 8-bit RGB image consists of 8-layers of binaries, or bipolar. In such way, each layer is as a single binary image for HNN. The results have shown the possibility and usefulness of HNN in RGB image recognition. Besides, the possibility of using wide number of RGB images stored in the net memory without sensed affection on the final results.

Quantum Grover Attack on the Simplified-AES

In this work, a quantum design for the Simplified-Advanced Encryption Standard (S-AES) algorithm is presented. Also, a quantum Grover attack is modeled on the proposed quantum S-AES. First, quantum circuits for the main components of S-AES in the finite field F2[x]/(x4 + x + 1), are constructed. Then, the constructed circuits are put together to form a quantum version of S-AES. A C-NOT synthesis is used to decompose some of the functions to reduce the number of the needed qubits. The quantum S-AES is integrated into a black-box queried by Grovers algorithm. A new approach is proposed to uniquely recover the secret key when Grover attack is applied. The entire work is simulated and tested on a quantum mechanics simulator. The complexity analysis shows that a block cipher can be designed as a quantum circuit with a polynomial cost. In addition, the secret key is recovered in quadratic speedup as promised by Grovers algorithm.

Quantum reversible circuit of AES-128

An explicit quantum design of AES-128 is presented in this paper. The design is structured to utilize the lowest number of qubits. First, the main components of AES-128 are designed as quantum circuits and then combined to construct the quantum version of AES-128. Some of the most efficient approaches in classical hardware implementations are adopted to construct the circuits of the multiplier and multiplicative inverse in