Revision history [back]

Since it is possible to do template matching with OpenCV, I started wondering.

Is it possible to do gender recognition with template matching?

My theory would be like this:

1. Create average faces of both genders from the training data
2. Isolate the face RoI from the test image, and resize said RoI to fit the average faces
3. Use template matching to match the test face RoI with the aforementioned average faces
4. Determine the gender by seeing which average face has the higher matching percentage

Is that feasible, and is it possible to do age recognition with a similar method, except with average faces from each age group?

Since it is possible to do template matching with OpenCV, I started wondering.

Is it possible to do gender recognition with template matching?

My theory would be like this:

1. Create average faces of both genders from the training data
2. Isolate the face RoI from the test image, and resize said RoI to fit the average faces
3. Use template matching to match the test face RoI with the aforementioned average faces
4. Determine the gender by seeing which average face has the higher matching percentage

Is that feasible, and is it possible to do age recognition with a similar method, except with average faces from each age group?group, and is template matching a time intensive task? Because I intend to do it in a live feed or a camera preview.

Since it is possible to do template matching with OpenCV, and after reading Kairos's blog post about it, I started wondering.

Is it possible to do gender recognition with template matching?

My theory would be like this:

1. Create average faces of both genders from the training data
2. Isolate the face RoI from the test image, and resize said RoI to fit the average faces
3. Use template matching to match the test face RoI with the aforementioned average faces
4. Determine the gender by seeing which average face has the higher matching percentage

Is that feasible, is it possible to do age recognition with a similar method, except with average faces from each age group, and is template matching a time intensive task? Because I intend to do it in a live feed or a camera preview.

Since it is possible to do template matching with OpenCV, and after reading Kairos's blog post about it, I started wondering.

Is it possible to do gender recognition with template matching?

My theory would be like this:

1. Create average faces of both genders from the training data
2. Isolate the face RoI from the test image, and resize said RoI to fit the average faces
3. Use template matching to match the test face RoI with the aforementioned average faces
4. Determine the gender by seeing which average face has the higher matching percentage

Is that feasible, is it possible to do age recognition with a similar method, except with average faces from each age group, and is template matching a time intensive task? Because I intend to do it in a live feed or a camera preview.