Virtual Mirror’s Documentation

Important

This documentation was generated on Jun 05, 2021, and is rebuilt with every master release.

Common Coloring Methods

apply_color(image, x, y, r, g, b, intensity)[source]

applies color to desired region

Parameters

  • arg1 (image) – input image

  • arg2 (int array) – X cordinates of the desired region that is going be colored

  • arg3 (int array) – Y cordinates of the desired region that is going be colored

  • arg4 (int) – red value of rgb color

  • arg5 (int) – green value of rgb color

  • arg6 (int) – blue value of rgb color

  • arg7 (float) – intensity of the applied color

Returns

the image with the applied color on the desire area

apply_blur(image, image2, x, y, gussiankernel, erosionkernel)[source]

applies blur on the desired region

Parameters

  • arg1 (image) – input image

  • arg2 (image) – input image with applied color on the desired region

  • arg3 (int array) – X cordinates of the desired region

  • arg4 (int array) – Y cordinates of the desired region

  • arg5 (int) – Gussian blur kernel

  • arg6 (int) – erosion kernel

  • arg7 (float) – intensity of the applied color

Returns

the image with the applied blur on the desired region

apply_blur_color(image, x, y, r, g, b, intensity, gussiankernel=51, erosionkernel=15)[source]

applies blur and color on the desired region

Parameters

  • arg1 (image) – input image

  • arg2 (int array) – X cordinates of the desired region

  • arg3 (int array) – Y cordinates of the desired region

  • arg4 – red value of rgb color

  • arg5 – green value of rgb color

  • arg6 – blue value of rgb color

Returns

the image with the applied clor and blur on the desired region

Note

this function is used when we need the color to be more faded on the edges of the region, this will give us a more natural look which is used on blush, concelaer and foundation.

moist(image, x, y, white)[source]

applies gloss texture on the lips

Parameters

  • arg1 (image) – input image

  • arg2 (int array) – X cordinates of the lips

  • arg3 (int array) – Y cordinates of the lips

  • arg4 – a value betwwen 0 to 255 (the higher this value th more white glossy texture will be )

Returns

the image with the applied gloss on the lips

Constants

This scripts holds the constant point indexes for each of the face components.

Utils

This script includes:

  • defenitions of functions for executing makeup workers in parallel to increase speed

  • functions to be called by the REST API, for enabling and disabling webcam, makeups and generating desired image.

class Makeup_Worker(bounds=None, instance=None)[source]

Defines a makeup worker instance

Parameters

  • arg1 (list) – list of the bounds related to the makeup worker, it’s assigned in normal worker functions and used in static worker functions

  • arg2 (class) – an instance of the related makeup worker, in case it exists (E.g. Lens)

class Globals[source]

Holds global values that are used by static workers and other functions

Properties:
cap

Video captured from webcam, can be closed and open upon request

face_detector

Instance of MediaPipe face_detection module

face_mesher

Instance of MediaPipe face_mesh module

makeup_workers

Dictionary of the makeup workers and their related information

makeup args

parameter names that are given to makeup workers, used in the api designed for this purpose

makeup instances

  • lens

  • eyeliner

  • eyeshadow

  • blush

  • lipstick

  • foundation

  • concealer

motion detection variables

  • prev_frame: holds the previos frame for comparing with the current frame

  • motion detected (bool): True if motion detected, False otherwise, starts with True

  • f_xmin: minimum x value among points of face crop

  • f_ymin: minimum y value among points of face crop

  • f_width: width of the face crop

  • f_height: height of the face crop

idx_to_coordinates

holds to coordinates calculated by turning face mesh index points to real coordinates in given image of face

output_frame

holds the given image without any change for using in static function or in special conditions

concealer_worker(image, r, g, b, intensity, out_queue) → None[source]

This function applies a concealer effect on an input image. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (list) – shared list for appending the output

concealer_worker_static(image, r, g, b, intensity, out_queue) → None[source]

This function applies a concealer effect on an input image with the calculated boundaries in the normal worker function. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (list) – shared list for appending the output

blush_worker(image, r, g, b, intensity, out_queue) → None[source]

This function applies a blush effect on an input image. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (list) – shared list for appending the output

blush_worker_static(image, r, g, b, intensity, out_queue) → None[source]

This function applies a blush effect on an input image with the calculated boundaries in the normal worker function. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (list) – shared list for appending the output

lipstick_worker(image, r, g, b, intensity, gloss, out_queue) → None[source]

This function applies a lipstick effect on an input image. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (list) – shared list for appending the output

lisptick_worker_static(image, r, g, b, intensity, gloss, out_queue) → None[source]

This function applies a lipstick effect on an input image with the calculated boundaries in the normal worker function. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (list) – shared list for appending the output

eyeshadow_worker(image, r, g, b, intensity, out_queue) → None[source]

This function applies an eyeshadow effect on an input image. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (list) – shared list for appending the output

eyeshadow_worker_static(image, r, g, b, intensity, out_queue) → None[source]

This function applies a eyeshadow effect on an input image with the calculated boundaries in the normal worker function. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (list) – shared list for appending the output

eyeliner_worker(image, r, g, b, intensity, out_queue) → None[source]

This function applies an eyeliner effect on an input image. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (list) – shared list for appending the output

eyeliner_worker_static(image, r, g, b, intensity, out_queue) → None[source]

This function applies a eyeliner effect on an input image with the calculated boundaries in the normal worker function. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (list) – shared list for appending the output

foundation_worker(image, r, g, b, intensity) → numpy.ndarray[source]

This function applies a foundation effect on an input image. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

Returns

The given image with the foundation effect applied on it, this image will be used as an input to other activated makeup workers in order for them to apply their effect on this image.

Return type

ndarray

foundation_worker_static(image, r, g, b, intensity) → numpy.ndarray[source]

This function applies a foundation effect on an input image with the calculated boundaries in the normal worker function. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

Returns

The given image with the foundation effect applied on it, this image will be used as an input to other activated makeup workers in order for them to apply their effect on this image.

Return type

ndarray

lens_worker(image, r, g, b, intensity) → numpy.ndarray[source]

This function applies a contact lens effect on an input image. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

Returns

The given image with the contact effect applied on it, this image will be used as an input to other activated makeup workers in order for them to apply their effect on this image.

Return type

ndarray

lens_worker_static(image, r, g, b, intensity) → numpy.ndarray[source]

This function applies a contact lens effect on an input image with the calculated boundaries in the normal worker function. This function is called by a threading function and its output is appended to a given list to be processed later.

Parameters

  • arg1 (ndarray) – input image, a crop of the face found in camera viewport

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

Returns

The given image with the contact lens effect applied on it, this image will be used as an input to other activated makeup workers in order for them to apply their effect on this image.

Return type

ndarray

join_makeup_workers(image) → numpy.ndarray[source]

Calls all the enabled makeup workers to produce their results on the given image. If enabled, lens and foundation workers are called beforehand and their results are given to other workers.

Parameters

arg1 (ndarray) – The face crop given to apply makeup filters on

Returns

The face crop with requested filters applied on

Return type

ndarray

How it works:

  1. Create a list to keep threads and another list to use as shared memory between the threads

  2. Check if foundation and lens worker are enabled, if so, apply them sequentially, this is to prevent merging issues with other workers’ results

  3. Iterate over makeup workers and start a thread for any one of them that is active, the arguments for each function is based on its own details. Also, the shared list is also passed to the target function

  4. Threads are added to the threads list and after creation, they are started and joined to finish with each other

  5. The shared list which now holds the results of all the worker functions is sorted by the index of each worker to create proper layering of filters

  6. Makeup filters are applied to the proper area one by one in the order of their predecence

join_makeup_workers_static(image) → numpy.ndarray[source]

Calls all the enabled makeup workers to produce their results on the given image without calculating the bounds. Only static methods are called in this function so the bounds that were calculated in the previous frame are used. If enabled, lens and foundation workers are called beforehand and their results are given to other workers.

Parameters

arg1 (ndarray) – The face crop given to apply makeup filters on

Returns

The face crop with requested filters applied on

Return type

ndarray

How it works:

The procedure is the same as join_makeup_workers(), with the exception that in this method, static makeup workers are called.

apply_makeup()[source]

Reads a frame from webcam camera and prepares it for the makeup workers to apply their filters on it.

Yields

byte array – The final image is encoded and converted to byte array so that sending it over a REST API is feasible.

How it works:
  1. Preparing the input

    1. A frame is read from the webcam into an image

    2. An unchanged copy of the image is saved into the global output_frame

    3. Image is flipped horizontally so that it is shown appropriately when displayed back to the user

  2. Motion detection

    1. A gray version of the image is generated

    2. If the global prev_frame has not been initilized, copy the gray image into it and continue

    3. Calculate the absolute difference between the current frame and the global prev_frame

    4. Threshold the difference to ignore minor differences

    5. Dilate the threshold result

    6. Find contours and calculate the contour area

    7. If the contour area is more than a certain amount, the global motion_detected flag will be set to True

  3. If motion has been detected:

    1. Process the image with MediaPipe’s face detection API, then crop the image with a padding to cover the whole face

    2. Process the cropped face image with MediaPipe’s face mesh API

    3. Convert the point indexes in src.cv.makeup.constants into actual coordinates in the face crop

    4. Call the joining function by passing the face crop image to it to apply the requested filters

    5. Placing the face crop in the original image

  4. If no motion has been detected:

    1. Crop the image using the boundaries calculated the last time we detected motion

    2. Call the static joining method by passing the face crop to it to apply the requested filters

  5. Sending the result:

    1. Encode the image in the desired extension (here we use .jpg)

    2. Put the image in a byte array and appending a proper header to be interpreted in the REST API`

enable_makeup(makeup_type='', r=0, g=0, b=0, intensity=0.7, gloss=False)[source]

Enables the requested makeup so that the joining methods will be applied it in the next frame. It also passes the received arguments from the API to the corresponding function.

Parameters

  • arg1 (string) – name of the makeup, it can be one of: [eyeshadow, lipstick, blush, concealer, foundation, eyeliner, lens]

  • arg2 (int) – rgb value of red color

  • arg3 (int) – rgb value of green color

  • arg4 (int) – rgb value of blue color

  • arg5 (float) – intensity of the applied makeup

  • arg6 (bool) – Determines whethere gloss needs to be applied or not (only for lipstick)

disable_makeup(makeup_type)[source]

Disables the requested makeup so that the joining methods will not be applied it in the next frame.

Parameters

arg1 (string) – name of the makeup, it can be one of: [eyeshadow, lipstick, blush, concealer, foundation, eyeliner, lens]

start_cam()[source]

Opens the camera with the desired camera index

stop_cam()[source]

Closes the camera when the user does not want to use the realtime virtual makeup functionality

Appendix

_images/absolute_diff.png

This picture shows the result of absolute differences calculated for a frame in motion detecion part of apply_makeup().

Indices and tables