Automatic attribution of ancient Roman imperial coins

Arandjelović, Ognjen

doi:10.1109/cvpr.2010.5539841

Cited by 48 publications

(72 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The first observation that emerges from the tables concerns the poor performance of the method which correctly recognized no more than 5% of query coins in the best case. This is consistent with previous reports in the literature [9], with the method showing any promise only in the context of the far simpler problem of coin instance recognition. The second clear observation concerns the superiority of the Hellinger distance based histogram metric as compared with the more conventional Euclidean distance.…”

Section: Resultssupporting

confidence: 82%

“…Alternatively, smaller values of k, more often used for object class recognition, can be seen as a means of generalization. Although this choice has not been explicitly discussed in the existing literature of automatic ancient coin analysis, implicitly the latter view seems to be dominant [9]. Considering the lack of systematic analysis to date, we take no a priori stance and instead conduct experiments for a range of values of k.…”

Section: Baseline Siftmentioning

confidence: 99%

“…None of these assumptions are realistic. The sole method which does not make this set of assumptions builds meta-features which combine local appearance descriptors with their geometric relationships [9]. Though much more successful than the alternatives, the performance of this method is still insufficiently good for most practical applications.…”

Section: Previous Workmentioning

confidence: 99%

“…In particular, all work to date has been highly unstructured and ad hoc in its evaluation methodology. Some authors use data sets with coins in different conditions and unstated distributions thereof [9], and others very small data sets with coins in extremely rare, museum grade [8]. Hence the current understanding of different methods' behaviour is not very well understood at all.…”

Section: Previous Workmentioning

confidence: 99%

“…This problem is of limited practical interest, its use being limited to such tasks as the identification of stolen coins or the detection of repeated entries in digital collections. Other works focus on coin type recognition, which is a far more difficult problem [7][8][9]. Most of these methods are local feature based, employing local feature descriptors such as SIFT [10] or SURF [11].…”

Section: Previous Workmentioning

confidence: 99%

See 4 more Smart Citations

Ancient Roman Coin Retrieval: A Systematic Examination of the Effects of Coin Grade

Fare

Arandjelović

2017

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Ancient coins are historical artefacts of great significance which attract the interest of scholars, and a large and growing number of amateur collectors. Computer vision based analysis and retrieval of ancient coins holds much promise in this realm, and has been the subject of an increasing amount of research. The present work is in great part motivated by the lack of systematic evaluation of the existing methods in the context of coin grade which is one of the key challenges both to humans and automatic methods. We describe a series of methods -some being adopted from previous work and others as extensions thereof -and perform the first thorough analysis to date.

show abstract

Section: Resultssupporting

confidence: 82%

Section: Baseline Siftmentioning

confidence: 99%

Section: Previous Workmentioning

confidence: 99%

Section: Previous Workmentioning

confidence: 99%

Section: Previous Workmentioning

confidence: 99%

See 3 more Smart Citations

Ancient Roman Coin Retrieval: A Systematic Examination of the Effects of Coin Grade

Fare

Arandjelović

2017

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

Understanding Ancient Coin Images

Cooper

Arandjelović

2019

Proceedings of the International Neural Networks Society

View full text Add to dashboard Cite

In recent years, a range of problems within the broad umbrella of automatic, computer vision based analysis of ancient coins has been attracting an increasing amount of attention. Notwithstanding this research effort, the results achieved by the state of the art in the published literature remain poor and far from sufficiently well performing for any practical purpose. In the present paper we present a series of contributions which we believe will benefit the interested community. Firstly, we explain that the approach of visual matching of coins, universally adopted in all existing published papers on the topic, is not of practical interest because the number of ancient coin types exceeds by far the number of those types which have been imaged, be it in digital form (e.g. online) or otherwise (traditional film, in print, etc.). Rather, we argue that the focus should be on the understanding of the semantic content of coins. Hence, we describe a novel method which uses real-world multimodal input to extract and associate semantic concepts with the correct coin images and then using a novel convolutional neural network learn the appearance of these concepts. Empirical evidence on a real-world and by far the largest data set of ancient coins, we demonstrate highly promising results.

show abstract

AnCoins: Image-Based Automated Identification of Ancient Coins Through Transfer Learning Approaches

Kiourt

Evangelidis

2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Automatic attribution of ancient Roman imperial coins

Cited by 48 publications

References 7 publications

Ancient Roman Coin Retrieval: A Systematic Examination of the Effects of Coin Grade

Ancient Roman Coin Retrieval: A Systematic Examination of the Effects of Coin Grade

Understanding Ancient Coin Images

AnCoins: Image-Based Automated Identification of Ancient Coins Through Transfer Learning Approaches

Contact Info

Product

Resources

About