Grab Motion
For marketing purpose, I built, designed and developed a desktop application capable of analyzing video files that are outputted from a GoPro camera. The cameras are installed into supermarkets while consumers walk and interact with products.
Later, the video folder paths of all days and all cameras are provided into the desktop application that recognizes motion according to defined areas and writes down the result as XML files and videos at the same time. This drastically reduces the human analysis of videos and avoids lots of operator work. The application synchronizes the cameras while the user is able to select the best take in each case. Finally, the result is delivered to marketing companies as processed data that is imported into third party software and visualized by specialists. It is within the scope of the project and roadmap to include features such as direction detection, gender, pedestrian count, face recognition among others. Much of its progress will depend on the success of the venue. Among other challenges, in order to achieve this project, I had to master in C++ language.
C++, OpenCV, Php, Bash, Electron, Python
https://josemanuelvigil.wixsite.com/grabmo
Marketing Motion Detector GoPro cameras output for analizing consummers
Region and camera selection
Project storage
Multi camera synchro with offline videos
Grabmo has been developed by Jose Vigil working in conjunction with partners and customers. Please send me a private message if you want to know more about the team and customers or find me on twitter at @josevigil.
Through the process of development, many of the challenges have been around handling GoPro output video format while doing multiple recognition jobs in parallel.
For this project, the Electron framework has been chosen to output the ongoing results on screen and progress. Due to its flexibility and communication with C++, it’s now possible to extend its functionality for multiple interactions with the user. The recognition part is based on C++ and was developed on top of the great Kerberosio, a security recognition framework owned by Cedric Verstraeten.
New challenges are in the roadmap such as direction detection, gender, pedestrian count, face recognition among others. Please contact me for any question or inquiry. Thanks very much for reading.
https://josemanuelvigil.wixsite.com/grabmo