International Journal of Signal and Imaging Systems Engineering (5 papers in press)

Regular Issues

• Design of high speed and low power multiplier using dual-mode square adder  
  by B. Jaya Lakshmi, R. Ramana Reddy, Naresh K. Darimireddy 
  Abstract: Adders are the basic building blocks in analogue and digital circuits for implementing arithmetic operations. Different adder designs are reported to obtain high speed, less area and low power dissipation. Dual mode logic (DML) and dual mode addition (DMADD) techniques can be used to achieve low power and high speed addition. Adders are the main blocks in multipliers. In this paper Braun multiplier is implemented using dual-mode (DM) square adder. The DM square adder architecture is a combination of DML and DMADD techniques. By incorporating the DM square adder in processors, power dissipation can be reduced. The full adder used in dual-mode square adder is static energy recovery full (SERF) adder which is faster and consumes less power compared to conventional full adder. For Braun multiplier using DM square adder power consumption is reduced by 63.54% and speed is increased by 90%. The proposed designs are implemented using mentor graphics tools in 130 nm technology.
  Keywords: dual mode logic; DML; dual mode addition; DMADD; DM square adder; SERF adder; Braun multiplier; low power.
  DOI: 10.1504/IJCAD.2019.10024523
   
  
• Syntactic approach to reconstruct simple and complex medical images  
  by Shilpa Rani, Kamlesh Lakhwani, Sandeep Kumar 
  Abstract: Pattern recognition is always a fascinating area of research for most of the researchers. A person can easily recognize the objects which are different in shape, size, color, and scale. Most of the available model uses a statistical approach for object recognition which is a good choice if noise is present in the image and images are simple but this method fails if patterns are more complex and there is a possibility of ambiguous results for complex pattern datasets. In that case, structural pattern recognition is more helpful. We focused on the syntactic approach for describing the features as knowledge and this technique represents the image in textual form using a syntactic approach which could be a great contribution in the field of theoretical computer science. Representation of image or object in textual form has been done through picture description language (PDL). There is a possibility that objects could be scattered in the image or maybe many objects are present in the same image. To identify the feature vector of all the objects, the gap-filling algorithm is applied which is a novel approach of the proposed method. The obtained feature vector can be used for the reconstruction of the original image. Experiments have performed on Brain MRI datasets and own dataset and the algorithm are able to convert a simple or complex image in textual form and reconstruction of an image using a knowledge vector is also done. To identify the performance of the reconstruction algorithm MAE, CPU time, and RMSE and iteration of the frame are calculated. MAE is 0.125 and 0.127 on its own and brain MRI dataset.CPU time is 1ms and 10ms on own and brain MRI dataset. Iteration of the frame is 2272it/s and 96it/s on own and brain MRI dataset. The performance of the algorithm is better than the existing techniques.GUI of the proposed work is also created for the better experience of user.
  Keywords: Pattern recognition; structural pattern recognition; PDL.
  
  
• Segmentation and detection of the retinal vascular network using fast filtering  
  by Nabila Rahmoune, Adel Rahmoune 
  Abstract: Changes in retinal blood vessels are a characteristic sign of many retinal diseases. Therefore, the automatic segmentation of vessels is an essential element for the diagnosis of different ocular diseases. In this paper, we present a novel algorithm for the detection and the segmentation of the vascular network of blood vessels in fundus images. Our algorithm employs two mean linear filters using the convolutional kernel, one directional along a line and the second on a square region, in combination with thresholding. The proposed approachs performance was tested on the public datasets DRIVE and STARE. Based on the test results, the mean segmentation accuracy, sensitivity, specificity and time complexity of retinal images in DRIVE are 94.27%, 97.01%, 66.20% and 1.63s and for the STARE database, they are 93.41%, 95.54%, 66.55% and 2.13s respectively. The proposed algorithm is simple and very fast. It achieved satisfactory mean segmentation accuracy with very low time complexity.
  Keywords: retinal blood vessel; image segmentation; mean linear filter; retinopathy.
  DOI: 10.1504/IJSISE.2022.10055616
   
  
• Image Correlation, Non-Uniformly Sampled Rotation Displacement Measurement Estimation  
  by Nicholas Wells, Chung See 
  Abstract: A rotation invariant image correlation algorithm is described which measures and locates the rotational error of individual spatial frequency components. This information may then be interpreted to track signal dependent signatures, analyze spatial frequencies with much higher bandwidths and form optimal matched and generalized correlation filters. The technique is based on image re-sampling and a non-uniform sampling interval that is adjusted depending on its distance from the origin of the polar map. A nearest-neighbor polar interpolated grid scheme, comparable to linear interpolation error, achieves accuracies of 0.1% of a degree. Preliminary measurements based on images containing natural and rigid structure are presented. The algorithm also has potential applications for data-driven image registration and deformation analysis with small variations.
  Keywords: image transforms; image analysis; correlation filters; rotation displacement measurement.
  
  
• Computational simulation of human fovea  
  by Prathibha Varghese, G.Arockia Selva Saroja 
  Abstract: Many metaheuristic algorithms have been developed with genuine inspiration from nature. From either of these photoreceptors through certain ganglion cells of such fovea towards the main cells of said visual cortex, every physical optical system is modeled in the form of cascading sub-filters. Commercially available optical sensors are nearly as photoreceptor-rich as the retinas of humans. But in terms of processing strategies and intensity, they still fall short of the human visual system. This idea has sparked research into the biological retina to better understand its information-processing capacities to copy the architecture to create mechanical visual sensors. Human fovea photoreceptor cones and rods have a hexagonal rather than a rectangular shape. Rectangular meshes are nevertheless frequently utilized since hardware implementation is straightforward. In contrast to rectilinear square meshes, hexagonal meshes offer a higher density of packing, continuous neighborhood connectivity, as well as improved angular rectification. If these camera sensors replicate the configuration of photoreceptor rods and cones as in the human fovea, computational efficiency to process a picture can also be increased. In that context, we provide a 2-D interpolation lattice conversion approach for creating hexagonal meshes, which is guaranteed to maintain alignment with our visual system and has a straightforward implementation and calculation process (HVS). This approach delivers a simulated hexagonal image for visual verification without needing a hexagonal capture or display device.
  Keywords: Hexagonal image; Spiral architecture; Hexagonal pixels; HVS; Square pixels.