Video-Image Segmentation : An Analysis Of Region-Based Image Segmentation

Flood is defined as “the temporary condition of surface water (lake, sea, rivers), in which the water level and/or discharge exceeds a certain value, thereby escaping from its normal confines” (Douben, 2006). Flood is one of the most threatening calamities that can occur to humanity. Urban areas in particular suffer from comparatively high flood risk due to high population number and density, multiple economic activities and lots of infrastructure and property values (Pelling, 2003).

Object detection in unorganized PCD. Nowadays, road features are becoming more complex, which leads having more complicated complaints in urban environments. Usually, PCD produces a wealthy set of data which need undergo a PCD process to identify and detect objects is in focus. It is important to extract objects, such as edges, pedestrians, curbs, and ends, from PCD. Over the last few years, many efforts have been made to detect objects, such as buildings, doors, etc. (Wang et al. 2014). Cluster analysis is one of the primary methods for enormous data analysis to detect objects, which is a massive PCD processing would help to recognize different natural grouping or structure. In the other word, the cluster can make a set of meaningful

Canthus is initially defined and treated as a corner (interest point). Interest point detectors like Harris and SUSAN are being studied and discussed in this report. These corner detectors are evaluated to find which one suits best for this application.

Zhiqiang Zhou et al. [24] proposed for long-distance wildfire smoke detection. While the long-distance wildfire smoke frequently moves slowly and lacks salient characteristics in the video, the detection is still a challenging issue. Unlike numerous traditional video based techniques that usually rely on the smoke colour or movement for initial smoke region segmentation, we apply the Maximally Stable Extremal Region (MSER) detection technique to remove local extremal regions of the smoke. This builds the initial segmentation of probable smoke region minimum dependent on the movement and colour data. Potential smoke regions are then chooses from all the probable regions with applying some static visual characteristics of the smoke, helping to reduce the non-smoke regions as several as possible. Once a potential smoke region is established, we keep tracking it by investigating the best-matched Extremal regions in the successive frames. At the same time, the propagating movements of the

Features are essential for any classification or analysis in image processing. There are many types of features which can be extracted from the images each gives its identical informations about the image. Here MA region supposed with properties such as shape, color and size which appears as a dark red colored circle shape. To identify MA and non- MA region feature vectors are formed for each candidate regions.

Floods are the most costly and damaging natural disaster in the world (Berz, 2000). The determination of streamflow (stage or discharge) at a river station plays an important role in environmental and water resources management. Flood routing, a basis for flood forecasting, is the process of determining progressively the timing and shape of a flood wave at successive points along a river reach. Approaches for flood routing have been based either on hydraulic (process approach) or hydrologic (conceptual or system approach). Hydrologic stream routing is based on the storage concept, and conversely, hydraulic channel routing is based on the principles of mass and momentum conservation. In predicting a particular hydrograph on its passage through a river reach, any flood modeling will involve a number of assumptions and simplifications. For many rivers, topographic information neither exit nor are readily available and consequently the use of dynamic models has some limitations (Barati, 2013). In addition, a close match between the observed flow rates and the computed ones is difficult to achieve with such techniques if some parameters used along a reach are not accurate. While high demands on computer resources, as well as on the quantity and quality of input data, restrict the efficiency of hydraulic models in practical applications, approximate models produce satisfactory results at considerably less expense with a limitation in their generality and accuracy depending on the

In vision detection system, the techniques used are either background subtraction based or feature detection based.

Model-based segmentation methods seek to transform this knowledge into intelligent algorithms that have prior knowledge about the structures of interest. These methods often consist of two steps. First, the location and the appearance of the primary model are initialized. Next, the shape and the appearance of the model are adjusted so that it is as close as possible to the dimensions measured in the image. An important feature considered for selecting a strategy when optimizing the shape of a model during segmentation is deformation. There are two main methods of deformation. One method addresses the deformation of the embedded space of the shape and the other is the modification of the parameters or degrees-of-freedom [15].

Abstract—This paper presents a comprehensive comparison of different high pass filtering techniques for Edge detection in both time domain and frequency domain. The paper examines various kernels and compares the efficiency of the filtering technique against the computation time for various sizes of images with various sizes of high-pass filter kernels. We have made use of Sobel filter as the standard filter kernel against which other techniques are compared.

To extract the small and large images by using the method of detecting edges of some predefined range of orientation in order to remove the problem related to detecting such boundaries.

Abstract- Interactive image segmentation has become more and more popular among researched in recent years. Interactive segmentation, as opposed to fully automatic one, supplies the user with means to incorporate his knowledge into the segmentation process. However, in most of the existing techniques, the suggested user interaction is not good enough since the user cannot intuitively force his knowledge into the tool or edit results easily. Therefore, in ambiguous situation, the user has to revert to tedious manual drawing. Presented method to develop as a combined segmentation and editing tool. It incorporates a simple user interface and a fast and reliable segmentation based on 1D segment matching. The user is required to click just a few "control points" on the desired object border, and let the algorithm complete the rest. The user can then edit the result by adding, removing and moving control points, where each interaction is follows by an automatic, real-time segmentation by the algorithm.

The simplest way to use image binarization is to choose a threshold value, and classify all pixels with values above this threshold as white, and all other pixels as black. Selecting proper threshold is very important task. In many cases, finding one threshold compatible to the entire image is very difficult, and in many cases even impossible. Therefore, adaptive image binarization is needed where an optimal threshold is chosen for each image area. Binarization is processing of converting color image in to binary image. In binarization, first we are converting color image in to Gray scale image using following formula. [2]There are various Binerization methods and in that various different algorithm used are as follows. Color image is converted into gray image and following algorithms are applied on gray scale image for converting it in to binary image.

Pattern recognition is a technique to differentiate different pattern into classes through the help of supervised or unsupervised technique. We have developed highly sophisticated skills for sensing the environment and taking actions according to what we observe. This sensing and understanding is mostly dependent on ability to differentiate between patterns. The pattern recognition ability if with the help of machine learning can be applied in machine. The machine ability to make decision like human being will be enhanced. Many applications such as data mining, web searching, face recognition etc has already been in uses which are based on the pattern recognitions. The objective of this review paper is to summarize and compare some of the well-known methods and application used in pattern recognition system.

The basic flow of algorithm is as show in Fig 3 As our aim is to do real time state of mind, detection of human so input image directly taken from webcam video. Therefore, we take few second video as input then extracting the frames from that video. After that number of frames, we apply some basic function on that image to improve the image quality. Colour image more complex for processing so that we convert tis colour image to the grey scale image.

Abstract—Saliency detection is used for many image processing applications. In this paper, single image saliency is used to improve the quality of the CT image. Single image saliency is obtained by measuring the contrast and spatial cue of a CT image. These cues are fused to obtain a kernel. In saliency filter the kernel is convolved with the CT image to improve the image quality for segmentation. The liver region is identified by adaptive thresholding and segmented by morphological operations and region growing method. The edge based active contour model using Distance Regularized Level Set Evolution is used to segment tumor from the liver region. The saliency based tumor segmentation provides better result in terms of accuracy.