摘要 :
In this work, we discuss the use of texture descriptors to perform writer verification and identification. We use a classification scheme based on dissimilarity representation, which has been successfully applied to verification p...
展开
In this work, we discuss the use of texture descriptors to perform writer verification and identification. We use a classification scheme based on dissimilarity representation, which has been successfully applied to verification problems. Besides assessing two texture descriptors (local binary patterns and local phase quantization), we also address important issues related to the dissimilarity representation, such as the impact of the number of references used for verification and identification, how the framework performs on the problem of writer identification, and how the dissimilarity-based approach compares to other feature-based strategies. In order to meet these objectives, we carry out experiments on two different datasets, the Brazilian forensic letters database and the 1AM database. Through a series of comprehensive experiments, we show that both LBP- and LPQ-based classifiers are able to surpass previous results reported in the literature for the verification problem by about 5 percentage points. For the identification problem, the proposed approach using LPQ features is able to achieve accuracies of 96.7% and 99.2% on the BFL and IAM and databases respectively.
收起
摘要 :
The paper presents a novel framework for the application of multiple features for handwritten data based identity recognition. Different types of features characterise different facets of the handwriting. We have designed a scheme...
展开
The paper presents a novel framework for the application of multiple features for handwritten data based identity recognition. Different types of features characterise different facets of the handwriting. We have designed a scheme for multiple feature based identity establishment using multi-kernel learning. A new formulation for multi-kernel learning using genetic algorithm has been presented. The efficacy of the framework using individual and combination of features is demonstrated for Devanagari script input.
收起
摘要 :
In the practice of forensic handwriting experts, the size of the writing on a questioned document may be different from that of known samples. In this study, the hypothesis of shape invariance of handwritten closed loops across si...
展开
In the practice of forensic handwriting experts, the size of the writing on a questioned document may be different from that of known samples. In this study, the hypothesis of shape invariance of handwritten closed loops across size increasing was tested. A Fourier methodology was applied to 2325 small letters (591 a loops, 547 d loops, 596 o loops and 591 q loops) and 692 enlarged letters (162 a loops, 173 d loops, 173 o loops and 184 q loops), in a population of 13 writers who were asked to write letters in their usual size and about three times larger. Most of the writers presented similar modifications when increasing the size of the loops; they produced enlarged loops significantly more round and less slanted towards the right or the left. Furthermore, a discrimination was demonstrated between the writers on the basis of the enlarged loops, with a correct classification rate superior to 90%, whatever the letter (a, d, o or q). A classification of the enlarged loops in their corresponding writer was then possible. On the contrary, when comparing the enlarged loops to the small ones, almost one half of the enlarged loops were allocated to a wrong writer. Shape invariance was thus not supported for this particular application. Consequently, when comparing documents with a different writing size, differences in loops shape should be interpreted cautiously because they may be due to a different writer, but they may also be due to an enlargement of the loops. Therefore, reference material of similar writing size to that of the questioned writing should be requested for the comparison of handwritten loops.
收起
摘要 :
Choosing ones words accurately makes for clear communication. Everyone knows this, yet my experience as an Associate Editor has shown me that writers, researchers, and others sometimes appear unaware of the implications of word ch...
展开
Choosing ones words accurately makes for clear communication. Everyone knows this, yet my experience as an Associate Editor has shown me that writers, researchers, and others sometimes appear unaware of the implications of word choice. It's worth remembering that words are not mere labels, but carry weight-the weight of history, social convention, perceptual habit, plus economic and political clout. We who practice medicine, do research, edit, and publish in The Journal of Alternative and Complementary Medicine need to be more sensitive to word choice. Some communication problems can be resolved with a simple change of word; others demand-and represent-a significant shift in attitude.
收起
摘要 :
Writer identification (writer-id), an important field in biometrics, aims to identify a writer by their handwriting. Identification in existing writer-id studies requires a complete document or text, limiting the scalability and f...
展开
Writer identification (writer-id), an important field in biometrics, aims to identify a writer by their handwriting. Identification in existing writer-id studies requires a complete document or text, limiting the scalability and flexibility of writer-id in realistic applications. To make the application of writer-id more practical (e.g., on mobile devices), we focus on a novel problem, letter-level online writer-id, which requires only a few trajectories of written letters as identification cues. Unlike text-\ document-based writer-id which has rich context for identification, there are much fewer clues to recognize an author from only a few single letters. A main challenge is that a person often writes a letter in different styles from time to time. We refer to this problem as the variance of online writing styles (Var-O-Styles). We address the Var-O-Styles in a capture-normalize-aggregate fashion: Firstly, we extract different features of a letter trajectory by a carefully designed multi-branch encoder, in an attempt to capture different online writing styles. Then we convert all these style features to a reference style feature domain by a novel normalization layer. Finally, we aggregate the normalized features by a hierarchical attention pooling (HAP), which fuses all the input letters with multiple writing styles into a compact feature vector. In addition, we also contribute a large-scale LEtter-level online wRiter IDentification dataset (LERID) for evaluation. Extensive comparative experiments demonstrate the effectiveness of the proposed framework.
收起
摘要 :
Problem statement: In recent years, fractal and multi-fractal analysis have been widely applied in many domains, especially in the field of image processing. In this study we presented a novel method for Persian and Arabic text-de...
展开
Problem statement: In recent years, fractal and multi-fractal analysis have been widely applied in many domains, especially in the field of image processing. In this study we presented a novel method for Persian and Arabic text-dependent writer identification based on fractal and multi-fractal features. Approach: From the images of Persian or Arabic words, we calculated their fractal dimensions by using the "Box-counting" method and then we calculated their multi-fractal dimensions by using the method of Diffusion Limited Aggregates (DLA). Results: To evaluate our method, we used 50 writers of the PERSIANDB database, each writer wrote 288 words (24 words repeated 12 times) with 2/3 of words are used for the learning phase and the rest is used for the identification Conclusion: The results obtained by using k nearest neighbor classifier, demonstrated the effectiveness of our proposed method.
收起
摘要 :
Diabetes was first extensively described by Aretaeus of Cappadocia, and his contemporary, Galen of Pergamum, in the second century ad. Aretaeus is said to have introduced the term diabetes, though there are some indications of pre...
展开
Diabetes was first extensively described by Aretaeus of Cappadocia, and his contemporary, Galen of Pergamum, in the second century ad. Aretaeus is said to have introduced the term diabetes, though there are some indications of previous references to the term. When referring to the disease, Galen accepts that the term belongs to ‘other writers’. There are, in fact, many other Graeco-Roman accounts of diabetes, and in this paper we also examine the texts of Pliny the Elder (first century ad), Rufus of Ephesos (early second century ad), Oribasius (fourth century ad), Stephanus Alexandrinus (fifth century ad), Aetius (sixth century ad), Alexander of Tralles (sixth century ad), and Theophilus Protospatharius (seventh century ad), together with his contemporaries Paulus of Aegina, Stephanus of Athens and Leon of Pella (ninth century ad). All these writers use almost the same terminology when referring to the disease, and consider that symptoms of excessive thirst and urination are due to disease of the kidneys and the bladder. Similar treatments are also advocated, and relate to nutritional advice, herbal medications, poultices, bloodletting and abstinence from diuretic substances. Although there are no notable differences between the writings of Byzantine physicians and the earlier descriptions attributed to Aretaeus and Galen, this literature testifies to the fact that these physicians were well aware of a disease that Galen considered rare.
收起
摘要 :
This exhibition seeks to re-introduce readers to a man who was once one of the most celebrated British writers. Now largely forgotten, Thackeray's name may be recognized only for his most famous novel, Vanity Fair. Thackeray's out...
展开
This exhibition seeks to re-introduce readers to a man who was once one of the most celebrated British writers. Now largely forgotten, Thackeray's name may be recognized only for his most famous novel, Vanity Fair. Thackeray's output, encompassing numerous novels and shorter works of fiction, scores of essays, articles, drawings, and other work, is much wider than Vanity Fair. In honor of the bicentenary of Thackeray's birth, this exhibition displays material from Houghton Library's substantial collection of Thackeray's books, manuscripts, correspondence, drawings, and other material. Together with items lent from private and institutional collections, they provide insight into Thackeray the devoted son, affectionate father, and loyal friend, in addition to the literary celebrity.
收起
摘要 :
Juozas Erlickas, known in Lithuania as Gvyvasis Klasikas, "The Living Classic," is a prolific Samogitian writer, satirist, political commentator, and musician; he brings to mind a weird mixture of America's Will Rogers, Mad Magazi...
展开
Juozas Erlickas, known in Lithuania as Gvyvasis Klasikas, "The Living Classic," is a prolific Samogitian writer, satirist, political commentator, and musician; he brings to mind a weird mixture of America's Will Rogers, Mad Magazine, and Idries Shah's Sufi tales. Very little of his work, however, has been translated, partly due to the difficulties inherent in his use of puns and word play, but more likely due to the fact that his writing doesn't fit neatly into any of the ordinary genres: it's really neither prose nor poesy – neither fish nor fowl.
收起