Steven:

In directory
	/group/project/VISION/web/3D_SKIN_DATA/
you will see some subdirectories DYYY, YYY=10/11/13/600/601/...
containing registered intensity and range images:
	pod1left_0X.jpg  pod1range_0X.tif
where X=1/2/3/4/...

  [the convention is slightly different in the 10/11/12 files but obvious]

The intensity images are the original skin view of size 2304x3456.
The range image is registered but is only 1/2 scale: 1152x1728.

I'll be adding some additional folders as time goes on.

I don't know how good the 3D recovery is for these. See task X below.

I don't know the disease classes for these images. I should be
able to get this eventually from Prof. Rees.

You should be able to load these into matlab the normal way with imread.

In folder

	/group/project/VISION/web/RAVINDRANATH

You can find the project code that Hriday Ravindranath wrote and left for
me. I'm not to sure what all is there, but there is some file names that
make some sence. I had a quick look at some of the matlab and it seemed
sensible (from a superficial look).

I think what you'll want to do is:
  1) look at each intensity image and identify a window around the spot that
	has the spot in the inner 1/2 of the window and smooth skin 
	surrounding it.
  2) make 2 new intensity and range images that are the same size (ie.
	range is not 1/2 scale) that cover only the window. Much smaller, 
	thus much more efficient to process. Make some consistent conventions
	on file names and types. Don't use lossy compression storage.
	You'll need to median filter each colour channel of 2x2 blocks
	to make them the same size as the range image.
  3) write a program that makes a new image that is cosine shaded of the
	range image. See me for details. Make cosine shaded images for 
	all range images
  4) Look at all range images. Choose the best for each folder if there
	is more than 1. Record folders that have bad 3D data.
  5) Use something like paint or gimp(?) to draw a binary mask image that
	has 1 for pixels inside the spot, 0 for normal skin
  6) look at Ravindranath's code and try to understand how the code relates 
	to the intensity features that he used. Produce variants for your
	project to do the same on the smaller images.

	...
   7) Develop the new features that we discussed
   8) Develop additional features
   9) Develop classifier & train (like IVR code)
   10) Develop feature selection algoritm


================================================================

some files are below
no certain diagnoses for these, just some notes
no certain quality on range images either


D10 - Epidermal\ cyst
D11 - Seborrhoic\ Keratosis.\ No\ match\ on\ Apex
D13 - Superficial\ spreading\ malignant\ melanoma
D65  - Congenital melanocytic naevi
D67 Seorrhoeic keratosis/

D600 - junctional naevus with mild atypia
D602 - ?
D603 - pilomatrixoma
D604 - ?
D605 - dermatofibroma
D606 - ?
D607 - blue naevus
D608 - moxoid cyst

D610 - intraeepidermal carcinoma
D611 - ?
D612 - ?
D613 - b cell lymphoma with plasmacytic differentiation
D614 - inflamed seborrhoetic keratosis
D615 - favours eczematous process
D616 - ?
D617 - ?
D618 - angioleiomyoma
D619 - eccreine poroma

D633 - PDT
D634 - PDT

D645 - ?
D646 - ?
D647 - seborrhoetic keratosis
D649 - seborrhoetic keratosis

D650 - simple lentigo
D651 - ?
D652 - multifocal superficial basal cell carcinomas
D653 - ?
D654 - lichen simplex chronicus
D655 - actinic keratosis
D656 - haemangioma
D657 - angiokeratoma
D658 - squamous keratosis
D659 - ?

D660 - seborrhoetic keratosis
D662 - ?
D663 - intraepidermal carcinoma
D665 - infiltrative basal cell carcinoma
D666 - ?
D667 - keloid scar
D668 - ?
D669 - nodular basal cell carcinoma

D670 - multifocal superficial basal cell carcinoma
D671 - ?
D672 - removed intraepidermal carcinoma
D673 - nodular and micronodular basal cell carcinoma
D674 - ?
D675 - lobular capillary haemangioma
D676 - keloid scar
D677 - ?
D678 - melignant melanoma - vertical growth phase
D679 - squamous cell carcinoma


15 AK - actinic keratosis
92 BCC - basal cell carcinoma
 1 Comedone      
 6 Dermatofibroma
 7 Haemangioma   
 8 Melanoma
56 Normal Moles
 4 Pyogenic granuloma
18 Seb K - seborrhoetic keratosis
22 SCC - squamous cell carcinoma