Colour comparison (1 Viewer)

Colour comparison.

Greetings,

I've been generating Hex colour codes that represent an average of a selected area of pixels from digital images of birds after reading this 1998 paper.. http://www.google.co.uk/url?sa=t&rc...uXrly_zZmM85GbwZLXV8fGQ&bvm=bv.53217764,d.d2k

For example.. I took 5 similar (lighting) good quality images of different Marsh Warblers and selected the pixels from just the bird's mantle. I then generated an average. My result was the Hex code 816A3A. I then did the same for Reed Warbler.. generating the Hex code BA8752

I then repeated the process using completely different set of Marsh and Reed warbler images and my result were codes very similar to the above.

I appreciate this is only a very simply example but the process could be greatly refined and so I'm wondering if such a methodology if developed might not aid and clarify colour terminology used in the identification of similar species and variation within taxa.

Your opinion would be very much appreciated.

I think the next thing you need is a blind test in which you apply your methodology to photos whose identity you don't know, and then see whether you identify them correctly using it. If you do, carry on, if you don't, back to the drawing board.

John

Good idea, but won't the shade of the colour vary with the amount of light? And with the exposure the camera decides on? I.e with a dark background, the camera might decide to brighten the whole image. If backlit, it might decide to darken it all.

And the recorded colour could vary with the colour balancing of the camera. Even if you've set it to a fixed value, what happens if the lighting is coloured, e.g a pink sunrise?

In practice, it might work though. What sort of variation did you get with your warblers?

John, at the Spurn migration festival I had a fascinating and immensely enjoyable time watching and listening to Clive McKay, one of the country's top ‘vis-miggers’.

Clive is doing some great research on the migration times of meadow pipits and is particularly interested in the movement of the Icelandic birds through Britain. I watched as he caught and ringed a number of mipits using the heligoland trap at the wire dump on Spurn. One of the ways to ID the Icelandic birds is that they would seem to have a subtle differences in feather colour, particularly over the mantle compared to the UK population... more olive, less golden brown values. The difference however is subtle. Apart from a general interest in the method I'm hoping that this technique might provide a useful analysis of that type of difference.

Yes Pshute.. illumination of the subject is a important factor but it was one the central concerns that Rafael Villafuerte and Juan Jose Negro addressed in the 1998 paper I provided a link to earlier. I think that they were onto something back then but that they didn't have the vastly more developed computer graphics programs and sophisticated digital image data that we have today.

pshute said:

What sort of variation did you get with your warblers?

Click to expand...

It's very early days in my exploration of all this but I'm currently doing an analysis of a larger collection of reed warbler images. One of the most immediate advantages of thinking about colour as a Hex code is that it encourages one to visualize a particular colour as a point in 3D colour sphere determined by the 3 major qualities of colour, the hue, saturation and the tonal value. (Hue at surface, the light to dark on an axis from pole to pole and the saturation a gradient towards the centre)

The reed warbler images I'm looking at not only have birds in a range of different illumination but also in terms of the freshness of the plumage. The more worn plumage while having colour on a similar tonal axis differ in saturation. I looking at developing some techniques to statistically analyse these different values.

Clive Hallam said:

John, at the Spurn migration festival I had a fascinating and immensely enjoyable time watching and listening to Clive McKay, one of the country's top ‘vis-miggers’.

Clive is doing some great research on the migration times of meadow pipits and is particularly interested in the movement of the Icelandic birds through Britain. I watched as he caught and ringed a number of mipits using the heligoland trap at the wire dump on Spurn. One of the ways to ID the Icelandic birds is that they would seem to have a subtle differences in feather colour, particularly over the mantle compared to the UK population... more olive, less golden brown values. The difference however is subtle. Apart from a general interest in the method I'm hoping that this technique might provide a useful analysis of that type of difference.

Click to expand...

In hand birds would also offer the best opportunity for standardising the light so obtaining the best possible results and enabling the research into very subtly different birds. I fully recognise the possibilities. I'm not quite so sure about the execution, which is why I think regular controlled checks such as blind tests to verify your progress must be included in your work. If you are onto something then you will pass the tests, if not, you will have to find a way to improve the technique, once you have analysed the reasons for failure.

Its very easy to write this from my armchair! Best of luck with your project.

John

Cheers John and for the advice, I appreciate your interest. :t:

Yes.. I think this method has particular potential in studying hand held birds or museum specimens where a "control card" can easily be included in the image. The difference between the control data and it's appearance in the picture could then be applied to plumage being examined to help standardize different illuminations.