Data Set Description For IMM
There are total 240 faces in the IMM data set.
All the data set are manually labeled with 58 points which be called the face landmarks point.
XM2VTS data set is the ideal face data set to testing your ASM/AAM fitting algorithm so far. IMM image size is 640x480. During our test, we use Viola Jones algorithm to detect the face. All 240 faces can be detected by 4 haarcascade_frontalface.xml files. The drawback of IMM data set is that the landmark points is too small.
Our ASM/AAM Fitting Result
We use xx_1x.jpg as the training set (40 faces). After ASM/AAM training, we fitting all the IMM data set (240 faces). Our best result is that the average point error is ave_err = 1.12369 for ASM fitting. Here is how we define the ave_err: for each face i (i=1..240 faces); do the ASM fitting, compare each point (58 points) with the manually landmarked point; calculate the average_error_i, i.e, sqrt(sum(calculated_value - manually_landmark_value)^2)/58; and then average the error, ave_err=sum(average_error)/240;
Our Face Top 5 Match Result
How to use AAM do the face match is an interesting topic.
Althrough there are some papers talk about using AAM do the face recognition, but we really don't see some good results.
Here is what we did:
Using AAM to extract each face feature which is a low dimension (only 38 data).
Then use the face feature data to do the matching. For each face in IMM, find the top 5 high similarity match.
The speed is very fast; It only cost 11ms to match one face against 240 faces.
The match rate is ok; It is 92.91% (i.e, 223/240, for each face, if we can find the same person excluding itself on the top 5, we take it as a success).
ASM/AAM Fitting Notation
- ASM1/xx-xx.jpg is the ASM fitting result with triangle mask.
- AAM1/xx_xx.jpg is the AAM fitting result with outline mask.
- AAM2/xx_xx.jpg is the AAM fitting result with triangle mask.
- AAM3/xx_xx.jpg is the AAM fitting result with appearance.
- AAM5/xx_xx.jpg is the AAM fitting result with appearance on the mean shape with outline.
- AAM7/xx_xx.jpg is the AAM fitting result with alignment to the mean shape with outline.
- AAM8/xx_xx.jpg is the AAM fitting result outline mask.
- AAM9/xx_xx.jpg is the AAM fitting result triangle mask.
