Final Movie - Everyone Has Lotus

Final Output - Everyone Has Lotus

Photo Shoot

I took a photo and video shoot in Room 505. It needed lots of stuff and efforts more than I expected. However, it was really fun and I learned a lot. Prepare more, better quality. Make a team and ask a help to friends was an important part that I learned.

Part 1# Video Projection Process

The room should be dark to project a video. to adjust a projector to the wall and model was also a really important part. 

What I used was a projector, a desk to put on the projector, a DVI-HDMI cable, a HDMI-C connector, a long power strip, and a Macbook. 

Of course, a model should be selected, too.

Part 2# Real-Time Face Detecting & Image Overlaying Process

The second part is a real-time face detecting & image overlaying. The room should not be dark such like part 1, because face detecting would be unstable if the room is too dark. There was a 50% light in the room. 

I changed my plan from projection to screen monitor because projection needs dark room and it makes face detecting unstable.

What I used was an Open CV, Python, a desk to put on my Mac, a DVI-HDMI cable, a HDMI-C connector, a long power strip, a stand, monitor, a chair, and a Macbook. 

The cable was shorter than I expected. It made me hard to capture widely. I cannot help but take a photo closer to the model.

Overlay Lotus on Eyes

Overlay Lotus on Eyes

Finally, I succeed to overlay lotus on eyes by tracking real time face with Open CV.

For the first time, I wanted to overlay lotus on the face, but it was not really cool.
It just looked like stick a giant sticker on the face.

So I decided to overlay lotus on eyes because it can show your lips, nose, and smile.
More fun~!:)

You can see the final output with face detectors signs below.

Overlay Lotus image on face

Overlay Lotus image on face

I overlayed lotus on my face.

It recognizes my face in real time and overlay and matches the lotus on my face automatically.

I don't know why but suddenly it stopped..

I applied to the recorded video and it still stopped.

I don't know why but it still does not work and occurs error.

I just resized the lotus and tested.
I need to fix some codes.

Use predictor to figure out the feature of face

Use predictor to figure out the feature of face

It worked not only for my face but also videos,

but it stopped when I started to using screen recorder..

Face recognition test work by Open CV [Python2.7 and Anaconda]

I used Open CV for face recognition.
It followed quite well even side face.
However, one code

face = faces[0]

sometimes occur error, and I don't know why because that code is just a simple code about how many faces will I track.