Saurabh Maheshwari

Fog removal techniques from images: A comparative review and future directions

Fog, haze and smoke are a big reason of road accidents. Fog reduces contrast level of the image that affects the visual quality of the image. In field of computer vision visual quality and visibility level of an image is affected by airlight and attenuation phenomena. Air particles, which present in atmosphere and affect the visibility level of any object, are called noise or unwanted signal between observer and object. For improving the visibility level of an image and reducing fog and noise various image enhancement methods are used. After enhancement is again restored the enhanced image by restoration methods. For improving the visibility level 4 major steps are used. First step is acquisition process of foggy images. Second is estimation process (estimate scattering phenomena, visibility level). Third is enhancement process (improve visibility level, reduce fog or noise level). Last step is restoration process (restore enhanced image). The main aim of this paper is to review state-of-art image enhancement and restoration methods for improving the quality and visibility level of an image which provide clear image in bad weather condition. We also compare prevalent approaches in this area through implementation of the methods keeping parameters common for critical analysis. In the end we provide the future scope for working directions in this area for the readers.

A review on vein biometric recognition using geometric pattern matching techniques

Due to increasing security issues, biometrics has gained attraction during past years in both academic and commercial domains. They range from physical traits such as retina, face, iris, fingerprint, etc. to behavioral traits such as signatures, gait, etc. Techniques in non-visible spectrum are based on infrared imagery namely near infrared, short wave infrared, mid wave infrared and long wave infrared. Among physical traits vein pattern prove to be more secured as blood vessels are underneath the skin and partially visible to naked human eyes, hence, it becomes hard for intruders to forge. Due to unique physiology of vasculature network, they are fit to be used for biometric authentication. A detailed description of vein recognition technology, arrangement, infrared imaging is presented. A well defined classification has also been provided for vein pattern extraction strategies. This paper is an approach to introduce interested researchers to the scope and methods in this field with detailed classification.

Permuted Image DCT Watermarking

This paper presents a novel watermarking technique for digital images using Pseudo Random(PN) Sequence. The host image is permuted by PN Sequence. The watermark is embedded into the mid band DCT coefficients of the image. The image is re-permuted using the same PN sequence to obtain the watermarked image. To ensure security, a PN sequence has been used which is known only to an authorized person. Extensive simulation results show that the proposed scheme can resist signal processing attacks such as median, average, motion and Gaussian filter as well as compression, blurring and sharpening of watermarked image. The proposed paper also analyzes the results obtained by varying the degree of permutation proving that the DCT along with permutation making the frequency domain watermarking arguably robust.

Foggy image enhancement using contrast limited adaptive histogram equalization of digitally filtered image: Performance improvement

Digital filters are used in signal processing to improve the quality of signal and to remove/reduce noise. Fog is a type of unwanted signal which create visuality difference between object and viewer. For removing this type of unwanted signal, digital filters are used. Digital filter is divided into two parts: finite impulse response (FIR) and infinite impulse response (IIR). The author introduced a new approach for improving the quality of foggy image and to increase peak signal to noise ratio. In this paper author used finite impulse response filter for foggy image which work over zero to finite time interval. In FIR filter, separable filter coefficients are used which reduce the amount of calculation. For improving image contrast hgamma correction function is used. After contrast adjustment of input image, contrast limited adaptive histogram equalization (CLAHE) is applied over an image and produce an enhanced image. The main aim of this paper is to improve the quality of foggy image using digital FIR filter and contrast limited adaptive histogram equalization. We also compare previous algorithms in this field through implementation of the methods keeping same parameter for critical analysis. In the end of this article author provide the future scope for working direction.

Robust Watermarking through Spatially Disjoint Transformations

An image watermarking scheme is introduced with an aim to satisfy the imperceptibility and robustness requirements. The image is divided into spatially disjoint areas. Watermarking is done by interleaving DCT (Discrete Cosine Transform), DWT (Discrete Wavelet Transform) and DFT (Discrete Fourier Transform). Multiple watermarks are simultaneously embedded in a single image in random order. Each area is selected sequentially and watermarked independently. The main advantage of this concurrent watermarking method is that it makes the watermark invariant to all the available frequency based attacks and geometric attacks. The shortcomings of one transform based watermarking scheme are overcome by the other. DFT based watermarking is invariant to geometric attacks and watermark detection attacks but it is not best suited for frequency based attacks like filtering and noise addition. Its shortcomings are overcome by applying the DCT and DWT based method in other areas which are resilient to frequency based attacks but fail against geometric attacks. Hence if any method fails against an attack the watermark can be efficiently extracted by the other implemented methods. Simulation results show that proposed method is able to withstand many image processing attacks.

Railway track breakage detection method using vibration estimating sensor network: A novel approach

Railway track security is prime concern of railways. Very few approaches have been implemented concerning the track breakage detection. In this paper we are proposing an algorithm for deployment of vibration sensors on the track and simulating communication among the sensors in real time. We also suggest a method to regularly update train engine about the tracks status and exact location of the track breakage. Communication takes place on a full duplex channel from sensors to sink and from sink to train. Multi-hop routing strategy has been used for which sender and receivers are sensors, trains and sink. The forwarding nodes are poles which have no sensor need, they are equipped only with transceivers. Certain advantages of the proposed method are: very few message parameters to maintain real time requirements in the decision making with highly consistent decision in comparison to previous approaches. Node has the minimum processing and decision making needs, so cost of nodes will be very less. Decision making is done separately for left and right track making the sensing more sensitive. No preprocessing on the data is done and exact output with precise coordinates, no fuzzy estimation. Aggregation in the proposed method is deterministic which gives exact location of crack whereas in complex fuzzy based approaches approximate location is identified so there is no guarantee for real time success.

Automated Brain Tumor Detection in Medical Brain Images and Clinical Parameters using Data Mining Techniques: A Review

mining is a growing field of research that intersects with many other fields such as Artificial Intelligent, Statistics, Visualization, Parallel Computing and Image Processing. Data mining techniques are good for Brain MRI image classification that can diagnose brain tumor and other diseases. In this paper we present an overview of the current research being carried out using the data mining techniques for the diagnosis of brain tumor. The goal of this study is to identify the most well performing data mining algorithms used on medical brain MRI and Clinical parameters. The following algorithms have been identified: Decision Trees, Support Vector Machine, Artificial Neural Networks and their Multilayer Perceptron model, and Fuzzy C-Means. Analyses show that it is very difficult to name a single data mining algorithm as the most suitable for the brain tumor detection or classification. At times some algorithms perform better than others, but there are cases when the properties of some of the above mentioned algorithm are combined together, they provide effective result. This paper also provides a critical evaluation of the literature reviewed, which reveals new facets of.

Railway Security System based on Wireless Sensor Networks: State of the Art

Railways are large infrastructures and are the prime mode of transportation in many countries. The railways have become a prime means of transportation owing to their capacity, speed, and reliability. Even a small improvement in performance of railways has significant economic benefits to rail industry. Thus, a proper maintenance strategy is required to govern optimization of inspection frequency and/or improvement in skill and efficiency. Accidents happening due to track breaking have been a big problem for railways for life security and timely management of services. This breakage needs to be identified in real time before a train actually comes near to the broken track and get subjected to an accident. In this paper, different kinds of rail defects inspection and maintenance methods are described and a basic algorithm is readdressed that makes use of wireless acoustic sensors for detecting cracks and breakages in the railway tracks.

Multi-domain Image Enhancement of Foggy Images Using Contrast Limited Adaptive Histogram Equalization Method

Enhancement process is used in image processing for improving quality and contrast of noise images or video. In digital signal processing, enhancement process is applied in spatial domain and frequency domain. In these two domains, process is applied over a pixel value but in frequency domain enhancement process is applied over frequency. In image processing for improving the quality of image-used image filtering, morphological operation, deblurring, ROI-based processing, neighborhood, and block processing. Contrast adjustment, morphological operation, ROI-based processing and deblurring enhancement processes are used in spatial domain. Digital filter is used in frequency domain. In this article, the author used top-hat filter for enhancement process and CLAHE methods are used in spatial and frequency domains. Contrast limited adaptive histogram equalization method is used in spatial domain and finite impulse response filter is used in frequency domain. At the end of this paper is shown the comparative results which prove frequency domain enhancement method with spatial domain, which is better than the spatial domain enhancement method.

Energy efficient railway track security using vibration sensing network

Accidents occurring in railway transportation systems cost a large number of lives. Many people die and several others get physical and mentally injured. There is certain need of advanced and robust techniques that can not only prevent these accidents but also eradicate all possibilities of their occurrence. We are proposing a sensor network which continuously monitors the railway track through the acoustic sensors and detect any discontinuity in the track. The sensor nodes are equipped with the acoustic sensors that can sense the vibration in the railway track due a coming train. They also estimate the theoretical value of vibration knowing their distance from the train. They compare sensed and estimated vibration intensities. If there is a major difference between these intensities then the track is suspected to be broken. The geographical positioning sensors are placed on the trains as well as on sensors. This complete process is real time in nature and should meet the timing deadlines. The nodes placed on the track and the forwarding poles are expected to work regularly for a long period of time without human intervention. These are generally battery operated so should be following some energy efficiency algorithm also. This paper is an attempt to optimize the energy used by the sensor nodes by using proposed intelligent sensor and lazy pole strategy. Very few approaches have been implemented concerning the track breakage detection. The sensors follow the proposed intelligent sensing technique whereas the poles follow the lazy poles strategy to make the approach energy efficient.