Testing recognition with cropped class images

Procedure

1. Use the cropped, non-smiling students (in class_nonsmiling_cropped) to compute 10 eigenfaces.  Show the average face and eigenfaces.

Eigenface 1	Eigenface 2	Eigenface 3	Eigenface 4	Eigenface 5	Eigenface 6
Eigenface 7	Eigenface 8	Eigenface 9	Eigenface 10	Average face

2. Have the program recognize the cropped, smiling students. You should expect only about 48% (10/21) accuracy with 10 eigenfaces.

a. Experiment with the number of eigenfaces used.  Try using the mean face plus 1 through 21 eigenfaces, at a granularity of 2

b. Plot the number of faces correctly recognized versus the number of eigenfaces used.

Questions

1. Describe the trends you see in your plots.   Discuss the tradeoffs; how many eigenfaces should one use?  Is there a clear answer?

With an increasing number of eigen faces used to compute the eigen space, the number of matches increases, but so does the computational time. In fact, it seems that the number of matches don’t increase as rapidly with the number of eigen faces used to compute the face space after a certain value. It is thus optimum in this case to use about 10 eigen faces.  However it is difficult to determine this without plotting a graph and noting the trends.

2. You likely saw some recognition errors in step 2; show images of a couple.  How reasonable were the mistakes?  Did the correct answer at least appear highly in the sorted results?  

Here are two mistakes when program was run with 21 eigen faces:

14	Match for 14 is non_smiling 7
17	Match for 17 is non_smiling 8

The mistakes seem reasonable considering the tilt of the heads and the dark regions being approximately in the same places.

For smiling photo 14, its non_smiling match is ranked 8^th and for smiling photo 17, its non_smiling match is ranked 5^th.

Cropping and finding faces

Procedure

1. Use your program to crop the elf.tga image. Use min_scale,max_scale, step parameters of .45, .55, .01.

2. Find a digital picture of yourself; if you really don't have one, use any portrait on the web.  Use your program to crop the picture.

3. Find the faces in two different photos (use the crop=false option)

a. First, try the group1.tga image given; you should be able to find all 3 faces.

b. Find and try another group photo.  This could be of family and/or friends, or one from the web.

Questions

1. What min_scale, max_scale, and scale step did you use for each image?  

For each image: min_scale = 0.45; max_scale = 0.55; step = 0.1

2. Did your attempt to find faces result in any false positives and/or false negatives?  Discuss each mistake, and why you think they might have occurred.

There were no false positives or false negatives in the pictures I used; all of which are shown above