Enhancement and Retrieval of Historic Inscription Images
01 Nov 2014-pp 529-541
TL;DR: By separating the text layer from the non-text layer using the proposed cumulants based Blind Source Extraction method, and store them in a digital library with their corresponding historic information, these images are retrieved from database using image search based on Bag-of-Words(BoW) method.
Abstract: In this paper we have presented a technique for enhancement and retrieval of historic inscription images. Inscription images in general have no distinction between the text layer and background layer due to absence of color difference and possess highly correlated signals and noise; pertaining to which retrieval of such images using search based on feature matching returns inaccurate results. Hence, there is a need to first enhance the readability and then binarize the images to create a digital database for retrieval. Our technique provides a suitable method for the same, by separating the text layer from the non-text layer using the proposed cumulants based Blind Source Extraction(BSE) method, and store them in a digital library with their corresponding historic information. These images are retrieved from database using image search based on Bag-of-Words(BoW) method.
Citations
More filters
01 Feb 2018
TL;DR: This model consists phase congruency and of Gaussian model based background elimination using expectation maximization(EM) algorithm, preprocessing and binarization, which removes the background noise completely where foreground characters are untouched.
Abstract: Epigraphs are important sources for reshaping our culture and history. They have a remarkable importance to mankind. But modern epigraphists find it difficult to interpret the information in scripts. It is mainly because inscriptions are eroded over a period of time due to natural calamities. Scripts of ancient times are largely unknown. Character sets used have changed from one form to another over the centuries. Therefore, for reading ancient scripts the characters have to be extracted. In this paper, a model for enhancement and binarization of historical epigraphs is proposed. This model consists phase congruency and of Gaussian model based background elimination using expectation maximization(EM) algorithm, preprocessing and binarization. In binarization, phase based features are used with specialised filters. Adaptive Gaussian filters are used to smoothen the output images. Weighted mean angle is calculated to differentiate the foreground from the background. EM algorithm removes the background noise completely where foreground characters are untouched. Proposed method is tested on different datasets of inscriptions and epigraphs. Obtained results are compared with the existing classical algorithms.
4 citations
Cites methods from "Enhancement and Retrieval of Histor..."
...In [2], proposes a method for the foreground extraction of text using cumulants based Blind Source Extraction(BSE) method, and store them in a digital library....
[...]
01 Nov 2015
TL;DR: A new technique to convert images of stone with inscriptions to binary forms is presented and it is compared to other existing algorithms for binary conversion and the results of the proposed method are better than the competing algorithms.
Abstract: Conversion of stone images with inscription to binary form pose several challenges due to perspective distortion, minimal distinction between foreground and background, lack of standardization of the text size and shape, less discernible color difference between the foreground and background and highly correlated noises. Traditional techniques to convert these images to binary form fail to handle all these challenges efficiently. In this paper we present a new technique to convert images of stone with inscriptions to binary forms. Initially, the Fast ICA (Independent Component Analysis) technique is used to enhance the minimal difference between text and non-text regions by decreasing the correlated noises. Then, the image is normalized using linear regression method. Cumulative residual entropy is obtained from the normalized image and it is analyzed to get threshold value to be used for conversion to binary image. The proposed technique is tested on images collected from world heritage site, "Hampi" and compared to other existing algorithms for binary conversion. The results of the proposed method are better than the competing algorithms.
3 citations
Cites background from "Enhancement and Retrieval of Histor..."
...Recently, enhancement and retrieval of historical inscription images is reported in [10]....
[...]
31 Jan 2017
TL;DR: A novel method to address the problem of enhancement and binarization of degraded manuscript images that applies a dual filtering technique for noise removal, Gaussian based adaptive thresholding technique and post processing using morphological operations that enhances readability of manuscript images is proposed.
Abstract: Manuscripts in physical form are easily damaged over time resulting in loss of important information. Hence, there is a need to preserve the knowledge these manuscripts hold by enhancing the readability of the damaged manuscript by applying various image analysis techniques and storing them in digital form to prevent further deterioration of manuscript information. There are multiple well-known methods for document enhancement but, they are not suitable for use in enhancing damaged manuscript images. We propose a novel method to address the problem of enhancement and binarization of degraded manuscript images that applies a dual filtering technique for noise removal, Gaussian based adaptive thresholding technique and post processing using morphological operations that enhances readability of manuscript images. Our method showed good performance on qualitative as well as quantitative evaluation performed on 27 digital manuscript images with uniform character formation and an overall pseudo F-measure of 60.12%. Furthermore, our method is also compared with other well-known document enhancing techniques to establish the better applicability of our technique to preservation of manuscript information.
Cites methods from "Enhancement and Retrieval of Histor..."
...Tomar [17] illustrated a technique to extract text from historic images by considering the problem as a blind source separation....
[...]
01 Mar 2019
TL;DR: Noise reduction techniques are applied on Estampage images, which are the exact replica of inscriptions and are the source of the authors' ancient history to preserve and processing for future.
Abstract: Noise reduction has an impact on preprocessing and quality enhancement of images. In this paper noise reduction techniques are applied on Estampages. Estampages are the exact replica of inscriptions and are the source of our ancient history. In preserving and processing for future, image processing techniques are applied. Noise reduction is mainly accomplished using filters, in our study various filters like Median, Gaussian, Wiener, Gabor, Box filtering, order statistics, morphological operations are applied and comparative study is made with respect to Structural similarity index. Suitable filters are combined with morphological operations to draw inference. Hybrid systems for noise reduction having Erosion combined with Wiener/Median followed by dilation processes are applied to test Estampage images. Visible and comparative studies using SSIM have evidenced good result for Erosion-Median-Dilation & Erosion-Wiener-Dilation combinations.
Cites methods from "Enhancement and Retrieval of Histor..."
...In [5], the bag of words procedure is used to retrieve the information about the input image from the database....
[...]
References
More filters
TL;DR: This paper presents a method for extracting distinctive invariant features from images that can be used to perform reliable matching between different views of an object or scene and can robustly identify objects among clutter and occlusion while achieving near real-time performance.
Abstract: This paper presents a method for extracting distinctive invariant features from images that can be used to perform reliable matching between different views of an object or scene. The features are invariant to image scale and rotation, and are shown to provide robust matching across a substantial range of affine distortion, change in 3D viewpoint, addition of noise, and change in illumination. The features are highly distinctive, in the sense that a single feature can be correctly matched with high probability against a large database of features from many images. This paper also describes an approach to using these features for object recognition. The recognition proceeds by matching individual features to a database of features from known objects using a fast nearest-neighbor algorithm, followed by a Hough transform to identify clusters belonging to a single object, and finally performing verification through least-squares solution for consistent pose parameters. This approach to recognition can robustly identify objects among clutter and occlusion while achieving near real-time performance.
46,906 citations
37,017 citations
"Enhancement and Retrieval of Histor..." refers methods in this paper
...After the extraction of three layers, the text information contained in the foreground layer is extracted by further processing which involves local thresholding through Otsu’s method [14] , morphological operations and median filter for smoothening purpose and finally a binary image is obtained which is free from any noise or unwanted background information and contain maximum text content....
[...]
...This layer is then binarized calculating a suitable local threshold level as per Otsu’s method proposed in [14]....
[...]
01 Jan 2011
TL;DR: The Scale-Invariant Feature Transform (or SIFT) algorithm is a highly robust method to extract and consequently match distinctive invariant features from images that can then be used to reliably match objects in diering images.
Abstract: The Scale-Invariant Feature Transform (or SIFT) algorithm is a highly robust method to extract and consequently match distinctive invariant features from images. These features can then be used to reliably match objects in diering images. The algorithm was rst proposed by Lowe [12] and further developed to increase performance resulting in the classic paper [13] that served as foundation for SIFT which has played an important role in robotic and machine vision in the past decade.
14,708 citations
"Enhancement and Retrieval of Histor..." refers background or methods in this paper
...A bag of words is a sparse vector which contains occurrence counts of local image features; in other words, BoW is a sparse histogram over a vocabulary where vocabulary is a set of discrete visual features such as Scale Invariant Feature Transform (SIFT) [11] descriptors....
[...]
...On the obtained digitized image, BoW descriptors are computed based on SIFT features [11]....
[...]
...1 Obtaining a Set of Bag-of-Words We take the set of digitized images from the database and obtain SIFT [11] descriptors based on SIFT feature points that are extracted from each of the image in the database....
[...]
...Secondly, these binary images are stored in a database and BoW descriptors based on SIFT features [11] of each image are computed and stored, which are used later for image retrieval from the database that are same or similar to the query image....
[...]
...scaling, translation, rotation and illumination changes [11]....
[...]
Book•
18 May 2001TL;DR: Independent component analysis as mentioned in this paper is a statistical generative model based on sparse coding, which is basically a proper probabilistic formulation of the ideas underpinning sparse coding and can be interpreted as providing a Bayesian prior.
Abstract: In this chapter, we discuss a statistical generative model called independent component analysis. It is basically a proper probabilistic formulation of the ideas underpinning sparse coding. It shows how sparse coding can be interpreted as providing a Bayesian prior, and answers some questions which were not properly answered in the sparse coding framework.
8,333 citations
13 Oct 2003
TL;DR: An approach to object and scene retrieval which searches for and localizes all the occurrences of a user outlined object in a video, represented by a set of viewpoint invariant region descriptors so that recognition can proceed successfully despite changes in viewpoint, illumination and partial occlusion.
Abstract: We describe an approach to object and scene retrieval which searches for and localizes all the occurrences of a user outlined object in a video. The object is represented by a set of viewpoint invariant region descriptors so that recognition can proceed successfully despite changes in viewpoint, illumination and partial occlusion. The temporal continuity of the video within a shot is used to track the regions in order to reject unstable regions and reduce the effects of noise in the descriptors. The analogy with text retrieval is in the implementation where matches on descriptors are pre-computed (using vector quantization), and inverted file systems and document rankings are used. The result is that retrieved is immediate, returning a ranked list of key frames/shots in the manner of Google. The method is illustrated for matching in two full length feature films.
6,938 citations
"Enhancement and Retrieval of Histor..." refers methods in this paper
...Bagof-Words (BoW) representation of a images for comparison has been a popular method and has shown excellent results in retrieval of word images [12] and videos [13]....
[...]
...Further, these feature descriptors are clustered using Kmeans algorithm into a defined number of bags and are trained, thus descretizing the descriptor space [12] [13]....
[...]
...Bag-of-Words was originally used in text classification and retrieval which has been to extended to use in retrieval of images and videos [12] [13] and has shown excellent results....
[...]