Handwritten character recognition has various potential in the field of document imageprocessing. It is one of the important aspects for systems like handwritten optical character recognizer, writer identification/ve...
详细信息
ISBN:
(纸本)9781467385640
Handwritten character recognition has various potential in the field of document imageprocessing. It is one of the important aspects for systems like handwritten optical character recognizer, writer identification/verification, automatic document sorter etc. In Bangla only few attempts are made towards character recognition. In this current study a relatively new attempt is made towards finding the dependency of writer information on character recognition by varying the inputs. this study will provide a better understanding of the input data for character recognition. Also it will help to know the Bangla characters better for writer identification/verification. Here, highest accuracy of 100% is achieved in case of numeral 7 applying LibSVM classifier.
the performance of an Infrared imaging system is often measured by its Detection, Recognition and Identification (DRI) ranges. In thermal cameras, acquired images will be of high dynamic range (typically 15 bits) and ...
详细信息
ISBN:
(纸本)9781467385640
the performance of an Infrared imaging system is often measured by its Detection, Recognition and Identification (DRI) ranges. In thermal cameras, acquired images will be of high dynamic range (typically 15 bits) and lack adequate contrast. An Indium Antimonide (InSb) detector operating in 1-5 micron wavelength has responsivity of the order of 30m V /degrees K resulting in very low differential pixel values even after non-uniformity correction. Mapping the grey levels of high dynamic range thermal data (typically 15 bits) to invariably less dynamic range displays (typically 8 bits), without loss of information, is a challenging task. Hence, image enhancement technique that preserves all the information content and displays with proper contrast is inevitable. Conventional enhancement techniques like AGC, Histogram Equalization and their many variants have the tendency of generating excess noise and assigning more display dynamic range to the dominating temperature data while suppressing the non-dominating temperature data. In this paper, a new enhancement technique of modified Adaptive Histogram Equalization (AHE) is proposed with its real time implementation on FPGA for high dynamic range thermal data. Algorithm aims to retain information content of high dynamic range thermal data, with improved contrast, low noise while compressing the dynamic range to suit the display resolution.
In this paper, an uncompressed domain video watermarking scheme resilient to temporal adaptation is proposed for scalable video coding. In the proposed scheme, each temporal layer has been separately embedded with a d...
详细信息
ISBN:
(纸本)9781467385640
In this paper, an uncompressed domain video watermarking scheme resilient to temporal adaptation is proposed for scalable video coding. In the proposed scheme, each temporal layer has been separately embedded with a different watermark which is generated by DCT domain decomposition of a single watermark image. A zigzag sequence of block wise DCT coefficients of the watermark image is partitioned into non-overlapping sets and each set is embedded separately into different temporal layers. the base layer is embedded withthe first set of DCT coefficient (which includes DC coefficient of each block) and successive layers are embedded with successive nonoverlapping coefficient sets. the coefficients of each set is chosen in such a fashion that uniform energy distribution across all temporal layers can be maintained. Experimental results show that the proposed scheme is robust against temporal scalability and robustness of the watermark increases withthe addition of successive enhancement layers.
Conventionally, High Dynamic Range (HDR) images are generated by fusing multiple exposure Low Dynamic Range (LDR) images, where the HDR output often suffers from artifacts due to misalignment of camera and presence of...
详细信息
ISBN:
(纸本)9781467385640
Conventionally, High Dynamic Range (HDR) images are generated by fusing multiple exposure Low Dynamic Range (LDR) images, where the HDR output often suffers from artifacts due to misalignment of camera and presence of dynamic objects in the scene. An efficient approach to overcome these issues is to use single shot HDR imaging. In this paper, we propose a method for generating an HDR image from a single LDR image. We first generate multiple exposures of the given scene using histogram separation by adopting varying bin sizes. the resulting LDR images are fused making use of the quality measures such as contrast, saturation and well - exposedness. the results show the effectiveness of the proposed approach which is verified qualitatively and in terms of various quantitative measures.
Automatic and reliable identification of pedestrians from multiple camera views is very important for video surveillance and can save a lot of manual effort. the significant variations in viewpoints, poses, illuminati...
详细信息
ISBN:
(纸本)9781467385640
Automatic and reliable identification of pedestrians from multiple camera views is very important for video surveillance and can save a lot of manual effort. the significant variations in viewpoints, poses, illumination and occlusions makes this problem very challenging. Most of the existing approaches addressing this problem handle drastic viewpoint change in a supervised way and thus require labelling new training data for a different pair of camera views. In this paper, we present a novel approach for pedestrian re-identification using stereo matching, which does not require any kind of training. the cost of the stereo matching of two images is used for evaluating the similarity of the images, without performing 3-D reconstruction. We show that this cost is robust to the large pose variations observed in the images captured from multiple cameras. the proposed pedestrian re-identification algorithm is built on top of a dynamic programming stereo matching algorithm. Experimental evaluation on the challenging VIPeR dataset shows the effectiveness of the proposed approach.
this paper proposes a new algorithm for restoration of gray scale images corrupted by salt and pepper noise(SPN). the proposed algoritm identifies a pixel as noisy if its intensity value is 0 or 255 and processes it u...
详细信息
ISBN:
(纸本)9781467385640
this paper proposes a new algorithm for restoration of gray scale images corrupted by salt and pepper noise(SPN). the proposed algoritm identifies a pixel as noisy if its intensity value is 0 or 255 and processes it using pixels in a 3 x 3 window. If the window consists of noisy and non-noisy pixels, then the pixel to be processed is replaced withthe trimmed median value of the non-noisy pixels. However, if only noisy pixels are there in the window then their mean value is used to process the pixel. the proposed method uses processed (i.e. the de-noised) pixels in the window while processingthe noisy pixels and shows significantly better performance, particularly at high noise density, as compared to various methods reported in literature. Experimental results show improvements both visually and quantitatively compared to other reported methods.
Steganography is the art of hiding secret data inside a carrier media. Most steganographic techniques suffer from the drawback that they are unable to retain the perceptual quality. Using saliency cues for developing ...
详细信息
ISBN:
(纸本)9781467385640
Steganography is the art of hiding secret data inside a carrier media. Most steganographic techniques suffer from the drawback that they are unable to retain the perceptual quality. Using saliency cues for developing an adaptive steganographic technique can help to alleviate this problem. In this work, a novel perception driven robust crypto-steganographic algorithm is proposed for embedding secure data in videos. the proposed scheme selects the payload regions based on natural scene statistics. To further strengthen the scheme and ensure intractability of secure data, the encrypted secret data is embedded in a random manner using jumbling sequence generator in the frames. We utilize perceptual hashing to evaluate the number of bit insertions that will not compromise the perceptual quality. A comprehensive performance evaluation of the proposed scheme is provided to detail the effectiveness. We demonstrate that the scheme shows a lot of promise in being robust against statistical and saliency based attacks.
We propose a novel approach of utilizing phenomic traits to automatically quantify stress in plants using machine learning techniques. Moisture deficit conditions cause change in leaf color due to decrease in chloroph...
详细信息
ISBN:
(纸本)9781467385640
We propose a novel approach of utilizing phenomic traits to automatically quantify stress in plants using machine learning techniques. Moisture deficit conditions cause change in leaf color due to decrease in chlorophyll content as chloroplast is damaged by active oxygen species. therefore, the proposed technique uses leaf color as the phenomic trait to assess stress levels using Relative water content (RWC) as a quantitative proxy. We extracted the change in leaf color in response to drought stress using the color features obtained using Random forest. A regressor has been modeled to predict the stress level of rice genotypes via RWC by employing colour histogram as a feature vector. the experiment was performed with pot images of different rice genotypes under normal and drought stressed conditions. We report a correlation coefficient of 0.89 obtained using this model demonstrating the capability of the presented technique for stress level predictions.
Stamps and logos are generally used for authenticating the source of a document. For automatic document processing, identification and segmentation of stamps and logos are essential. In the past, methods to detect sta...
详细信息
ISBN:
(纸本)9781467385640
Stamps and logos are generally used for authenticating the source of a document. For automatic document processing, identification and segmentation of stamps and logos are essential. In the past, methods to detect stamps and logos were limited to specific shapes, colors, or training data. However, stamps and logos can be of any shape or color. In this paper, we have proposed a novel stamp and logo detection technique. Our approach is based on the fact that stamps and logos, in general, are not the primary contents of a document. this fact motivates us to propose an outlier detection technique for the same purpose in a feature space. Based on some geometric features, the detected outliers are classified as stamps and logos. Our method shows good performance in case of separating them from text. Moreover, this technique is capable of detecting logos as well as chromatic and achromatic stamps.
Codebook model is a widely used method for segmenting foreground pixels. However, it often generates the erroneous detection results with a dynamic background. this paper proposed a multi-layer codebook model for segm...
详细信息
ISBN:
(纸本)9781467385640
Codebook model is a widely used method for segmenting foreground pixels. However, it often generates the erroneous detection results with a dynamic background. this paper proposed a multi-layer codebook model for segmenting foreground pixels and a separate layer is utilize to disparate static background from dynamic background. It improve processing speed as system keep history of uncovered background region. Proposed method reduces erroneous positive pixels detected conventionally as ghost region when pixel belongs to background suddenly start moving. To eliminate shadow/illumination effects cone shaped color distance map is utilized in lieu of cylindrical. During experimentation proposed method is tested over numerous videos with complex illumination and background situations. the experimental result shows improvement over rudimentary codebook model and other state-of-art background subtraction model by reducing erroneous positive pixels detected such as ghost region and enhance foreground detection.
暂无评论