Skip to content

Pavithra-Rajan/Beauty-Predictor

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 

Repository files navigation

Beauty-Predictor

The Chicago Face dataset has been used to create a Machine Learning model as well as a Deep Learning model with Convolutional Neural Network and transfer learning with the VGG dataset. The Chicago face dataset comprises images of around 597 males and females with facial proportions and an attractiveness score. So, I use this dataset to rate myself. The machine learning model created is purely regression-based, and it assesses faces based on the dimensions of parts of the face. The second model created utilized images of 597 people, and a dense convolution with the VGG face dataset for transfer learning. So, an image can be loaded on to this model, and it would return the score. This model basically rates us purely based on the image, while the one mentioned above has a mathematical approach.

The database can be downloaded from:

https://chicagofaces.org/default/ (In order, to access the data, a request form has to be submitted stating the purpose and the link will be sent as a mail.)

The VGG face-model from:

https://drive.google.com/file/d/1CPSeum3HpopfomUEK1gybeuIVoeJT_Eo/view?usp=sharing

Dimensions required by the ML model:

Nose_Width Distance between outer edges of nostrils at widest point.

Nose_Length Distance between Forehead Bridge at level of visible upper eye edge to nose tip. For the Nose Tip, imagine a spot where the nose protrudes from the face. Optional Tip: You can make a temporary line across the eyes to get a good visual reference of where the upper eye edge is on the bridge of the nose.

Face_Length Distance between hairline and base of chin. Note: For some people, there may be a double chin. You may have to guesstimate.

Avg_Eye_Height Average distance between upper and lower edge of visible eye within eyelids at center of pupil for right and left eye.

Avg_Eye_Width Average distance between inner and outer corner of eye for right and left eye.

Face_Width_Cheeks Distance between outer edges of cheeks at most prominent points.

Forehead Distance from center of hairline to the center between the eyes. Tip: You can make a temporary box that goes between the eyes to get a good visual reference on where to start on the bridge of the nose.

Asymmetry_pupil_top Absolute value difference between distance between pupil center to hairline for right and left.

Pupil_Lip_L Distance between pupil center to top edge of lips. Asymmetry_pupil_lip Absolute value difference between distance between pupil center to top edge of lips for right and left.

Midcheek_Chin_L Distance between midcheek and bottom of chin.

Cheeks_avg Average distance between midcheek and bottom of chin for right and left.

Midbrow_Hairline_L Distance between midbrow to hairline. This should be above the pupil of each eye in the middle of the eye brown.

Heartshapeness Face width at cheeks divided by face width at mouth

Noseshape Nose width divided by nose length

LipFullness Lip thickness divided by face length

EyeShape Eye height divided by eye width

UpperHeadLength Forehead divided by face length

MidfaceLength (Average pupil to lip for right and left) divided by face length

ChinLength Bottom of lip to chin divided by face length

ForeheadHeight (Average midbrow to hairline for right and left) divided by face length

CheekboneHeight (Average midcheeck to chin for right and left) divided by face length

CheekboneProminence (Face width at cheek minus face width at mouth) divided by face length

FaceRoundness Face width at mouth divided by face length

fWHR Facial width to height ratio

Each measurement taken in cms must be converted to inches (divide by 2.54) and multiplied by a factor of 100.

Releases

No releases published

Packages