DeepFace-Powered Image Attribute Analysis
Budget: $10 – $30 USD
Make Python script to scan folder of images for facial attribute analysis using DeepFace Open Source Facial Recognition Library (https://viso.ai/computer-vision/deepface/).
Python script location should be in same folder as all *.jpg files
User will run the script and when script is running it will append to file result.txt in the same folder.
The output format for result.txt should be *.csv, or comma delimited *.txt, and contain 4 fields:
image-file-name, age, gender, race
One per line in the output for each image scanned. emotion data not required. only display the race with the greatest value or "dominant_race"
If no face is found in image, then output should be: image-file-name, nothing, nothing, nothing
Script should run fast and smooth and be able to handle 750k+ images when run at 95%+ accuracy.
A sample of 5k images is provided for testing. Please send me the results of the sample to check and be awarded the project. After I review a sample, if it looks good, I will award you the project.
Also after awarding the project, I might need some help installing the packages in python.
Thank you
Here is the link for 5k sample images to test with:
https://www.dropbox.com/scl/fo/o2mekgo53brydm1mgof6m/AIdgyrtqaQUDTiJSvb02x_0?rlkey=aa4abv1l9gq7n9f8jghi47l57&st=oka4el8l&dl=0
The default VGG model that comes with DeepFace should be fine for the script. My budget is $20 for this project.
See attached quick guide
Python script location should be in same folder as all *.jpg files
User will run the script and when script is running it will append to file result.txt in the same folder.
The output format for result.txt should be *.csv, or comma delimited *.txt, and contain 4 fields:
image-file-name, age, gender, race
One per line in the output for each image scanned. emotion data not required. only display the race with the greatest value or "dominant_race"
If no face is found in image, then output should be: image-file-name, nothing, nothing, nothing
Script should run fast and smooth and be able to handle 750k+ images when run at 95%+ accuracy.
A sample of 5k images is provided for testing. Please send me the results of the sample to check and be awarded the project. After I review a sample, if it looks good, I will award you the project.
Also after awarding the project, I might need some help installing the packages in python.
Thank you
Here is the link for 5k sample images to test with:
https://www.dropbox.com/scl/fo/o2mekgo53brydm1mgof6m/AIdgyrtqaQUDTiJSvb02x_0?rlkey=aa4abv1l9gq7n9f8jghi47l57&st=oka4el8l&dl=0
The default VGG model that comes with DeepFace should be fine for the script. My budget is $20 for this project.
See attached quick guide
Related categories:
Python
Software Architecture
Image Processing
Computer Vision
Deep Learning
Facial Recognition