This is a 2 people project, where users can download a set of images from Flickr, then perform a object detection / image retrieval. It's built in C++ using CVUI, OpenCV and libcurl library.
Users can download recent images from Flickr. This is done by calling the Flickr API (flickr.photos.getRecent) with HTTP GET using the curl library. Then, the HTTP data received are then converted to JSON in order to extract the image’s direct URLs, which will be downloaded as a .jpg image using the curl library again.
Setting window
In the setting window, users can select the methods used to calculate the object
detection (pixel difference, histogram difference, Hough transform, SIFT), they can
also configure the scale parameter as well as the scale/location step for the search
windows.
Detection window
In the detection window, it will show the input image as well as the image that is being
searched on both sides of the window. A red square on the search image represents the
current search window. A difference value is shown on top where smaller means a higher
detection rate. The search progress is also shown as the progress bar at the bottom. There
is also a speed slider where users can adjust the search speed.
Result window
After the detection progress is complete, the result window is shown, users can click the
left/right button to browse through different results.
Creation date: November, 2021
My role in this project: Flickr downloader, object detection, UI
