Build an efficient system that would optimize the transportation of relevant products to the outlets where the possibility of sales would be more.
What it does
Capture the images from a camera installed in the premises of the outlet of the store to get demographic information like gender, age, etc. Based on this information, the businesses can improve their marketing campaign, targeted advertisements, and product placements.
How we built it
We used Deep Learning techniques incorporated with computer vision algorithms. We used pre-trained VGG face model and used the concept of transfer learning to predict age and gender from the image. We used Haar Cascade to detect the faces in a frame. We used Flask as a back-end to render html. Using jQuery distribution of data regarding the consumers was shown.
Challenges we ran into
The main challenge was to track the number of people in a given scene. Dealing with face recognition model was a little troublesome which hindered our progress for human face recognition and generating their vectorized latent space representation.
What's next for CustTrack
Cameras would generate heat-maps which would help in understanding footfall patterns in different commercial spaces.