We are getting access to camera using getUserMedia() using different MediaTrack constraints and settings and then take a picture using takePhoto(). As of now, we are displaying the image on a canvas. It would be useful to save the image somewhere (maybe filesystem, like WICG's nativefilesystem) and then we can browse through the gallery. For filters/ we display using cv.imshow(), so we need to explore cv2.imwrite('result.png',img) on browsers.

Similarly for loading images from gallery to camera app, atleast on Android we can use the File/Photo Picker or the Photo/Video Picker which Kenneth reported.