Music🎵 player with Emotion😂 😥 😡 😱 Recognition 
Sourcerer
Full Tutorial here
Code Requirements
- Tensorflow
- Download my repository
- Own Expression dataset(NOTE: You can downlaod expression images from google, or you can record your video make diffrent expression ,and converts into Grayscale images(For more accurate prediction))
- Song dataset
What steps you have to follow??
- Download my repository
- Make 'Images' folder in your project ,make subfolder for emotions like Happy,sad,Angry.
- Put
Face_crop.py&haarcascade_frontalface_alt.xmlin every type of image folder,ex : put this program in "happy' image folder and run this program it will detect faces from images and convert it into grayscale and make a new images in same folder. - Make 'Songs' folder make subfolders for emotions and put Songs,Like Happy songs in happy folder.
- After that you have to create model, for that copy code from code.txt file and open CMD in your project folder and paste it & enter
- It will take training aaround 20-25 minutes so keep patience.
- After training it will create two files
retrained_graph.pb&retrained_labels.txt - Now run
music_player_webcam.py(give proper path of songs and Mediaplayer according to your location in code) - If you want to fetch video from your mobile cam than use
music_player_android.py,but you have to install IPWebcam app in your system and replace your server URL with my URL - That's all
How it works? See:) FREE VERSION WITHOUT GUI
Video demo :- FREE VERSION
IMPORTANT if you want to purchase with GUI
- This is just coded version with no GUI, i ggiven it to opensource ,but if you want to purchase a GUI version ,then go to here STORE
Features of full GUI version
- Full GUI (Tkinter)
- Own developed Music player with all basic functionalities
- See here are some screenshots of full GUI version.
Screenshots
.png)
.png)
.png)
.png)
.png)
This guy purchased this project from me and won prize in Zhejiang university of technology China
Poster Presentation
Notes
- It will require high processing power(I have 8 GB RAM & 2 GB GC)
- If you think it will recognise expression just like humans,than leave it ,its not possible.
- Download 300 Images for every expression(you can use batch downloader)
- Noisy image can reduce your accuracy so quality of images matter.