To browse Academia. Skip to main content. Log In Sign Up. Tinne Tuytelaars. Basura Fernando.
Only recently this task has been addressed in computational image analysis perspective. The characteristics of the device used to acquire each photograph Fig. Some photographs from left to right in chronological are discriminative features for this task.
We aim at extracting order.
COLOR FEATURES FOR DATING HISTORICAL COLOR IMAGES Basura Fernando? y, Damien Muselet, Rahat Khan and Tinne Tuytelaars?y?y PSI-VISICS, KU Leuven, iMinds, Belgium Universit Jean Monnet, LaHC, Saint-Etienne, France yALCoV-ISIT, UMR CNRS / University of Auvergne, Clermont-Ferrand, France ABSTRACT Estimating the age of historical . Dating Historical Color Images. By Frank Palermo, James Hays and Alexei A. Efros. Abstract. Abstract. We introduce the task of automatically estimating the age of historical color photographs. We suggest features which attempt to capture temporally discriminative information based on the evolution of color imaging processes over time and Author: Frank Palermo, James Hays and Alexei A. Efros. CiteSeerX - Document Details (Isaac Councill, Lee Giles, Pradeep Teregowda): Abstract. We introduce the task of automatically estimating the age of historical color photographs. We suggest features which attempt to capture temporally discriminative information based on the evolution of color imaging processes over time and evaluate the performance of both these novel features .
Left photograph is from s next ones from s, such characteristics from a historical color photographs. The s, s and the last from s. This figure clearly acquisition device mainly effects two properties of the colors: shows the difficulty of the task. We propose two color features that take advantage of these observations. We show that these two color descriptors namely color deriva- tives and color angles attain the state-of-the-art in the context of image dating.
In this aim, Palermo et al. In particular, they exploit the probabil- derivatives, Color angles, Imaging process ity to have a certain color saturation for each considered hue.
The hue and saturation they are using are estimated from the 1. Fur- ous areas of studies such as history, genealogy and anthropol- thermore, we show that, given one image, while being very ogy.
They are interested in the conservation and preservation useful, its color distribution is not the most accurate feature of old photographs. More often than not this information is to discriminate between decades. Moreover, there is a growing interest to- wards cultural heritage projects such as Flickr Commons  that tries to make historical photography collections accessi- Unlike classical color descriptors, our proposed features ble online. These preservation projects have a real necessity are device-dependent.
We consider that the characteristics to estimate the age of a photograph automatically. Only re- of the devices used to acquire one photograph are very ac- cently, the task of automatically estimating the age of histori- curate features to date this photograph.
For example, the cal photographs has gained some attention [2, 3].
Therefore, we aim at extracting these characteristics photograph seems to be the most popular method at the mo- from each image. From the imaging process, we claim that ment. Mostly, these manual techniques rely on visual cues the acquisition devices impact two different properties of the such as particular fashion or hair styles worn by human sub- colors: first, the global distributions of the RGB derivatives jects [4, 5].
Dating Historical Color Images
In some specific instances, tive pixels along the edges. Thus, we propose to use two dif- estimating the age of the photograph can be done using do- ferent and complementary color features for the dating task, main specific knowledge such as in  where military histori- namely color derivatives and color angles introduced in sec- cal photographs are successfully dated using such techniques.
We show in section 4 that they perform significantly Following the work of Palermo et al. As explained in , each step of this acquisition process can be associated with The most related work for dating historical images is the one one specific and often unknown transform. In this paper, we from Palermo et al. They propose to split the available claim that some of these transforms are impacting the global historical images into discrete time interval classes and de- statistics of the colors in the resulting photographs while oth- fine the problem as an image classification problem.
While ers make the relation between scene radiance and pixel in- many definitions of these time intervals are possible, decade tensity nonlinear.
For example, the white balance and the e. A novel color descriptor called Condi- mapping, aesthetic effects, gamma correction or old film re- tional Probability of Saturation Given Hue is proposed.
This sponses are adding non-linearity to the mapping . Schindler et al. Color derivatives which requires many overlapping images of the same scene Palermo et al. Following a similar idea Xu et al. Then they deduce hue over many years. Recently, Rematas et al. In this paper, instead age of a man made object in an image using modern descrip- of using approximate transforms between color spaces, we tors such as SIFT and encoding methods such as Fisher vec- propose to exploit the variations of the color distributions over tors, while Lee et al.
These methods do not inspect the photo- Our motivation comes from the fact that RGB global color graph as a physical artifact but rather rely on object recogni- distributions have been shown to reveal information about tion techniques to identify specific patterns that are correlated the acquisition conditions such as the lights and the cam- with time or the decade.
Even though, these methods can be era [17, 18]. For example, that is not restricted to some kind of contents and that is based the average RGB components are used by the grey-world al- on two different and complementary color descriptors intro- gorithm and the convex-hull of the RGB distribution is the duced in the next section.
More recently, it has been shown that taking into account the 3D distribu- 3. We also define the task of estimating the age of a photo as The idea behind these algorithms is that RGB zeroth, first and determining the decade during which a historical color pho- second derivatives coarsely follow canonical distributions in tograph was taken.
Rare Historical Photos in Colour You Must See
Shrimpton, J. Countryside Books Google Scholar. Storey, N. Sulonen, P.
Taylor, M. Family Tree Books Google Scholar.
Dating historical color images
Torralba, A. Wilhelm, H. Preservation Publishing Company Google Scholar. Xiao, J.
Efros 1 1. Personalised recommendations.
Cite paper How to cite? ENW EndNote.
Dating Historical Color Images
Share paper. For the challenging classification task of sorting historical color images into the decade during which they were photographed, we demonstrate significantly greater accuracy than that shown by untrained humans on the same data set. Additionally, we apply the concept of data-driven camera response function estimation to historical color imagery, demonstrating its relevance to both the age estimation task and the popular application of imitating the appearance of vintage color photography.
Dating Historical Color Images People. Frank Palermo James Hays (Brown University) Alexei A. Efros. Abstract. We introduce the task of automatically estimating the age of historical color photographs. We suggest features which attempt to capture temporally discriminative information based on the evolution of color imaging processes over time.
The software developed during the course of this research is currently being prepared for release to the public. During this time, the link below will always point to the most current version of all source code which has been made publically available. Please check back periodically for ates.
butterfishny.com is a platform for academics to share research papers. Dating Historical Color Images 3 plications such as CameraBag , Hipstamatic , and Instagram . While the fundamental technique of manipulating color channel response curves is well known (e.g. ), we are unaware of any existing publication which has at-tempted to learn such models using a data-driven approach. Abstract. We introduce the task of automatically estimating the age of historical color photographs. We suggest features which attempt to capture temporally discriminative information based on the evolution of color imaging processes over time and evaluate the performance of both these novel features and existing features commonly utilized in other problem domains Cited by: