Converting a video to keyframes requires that the video compression (its Codec) is frame-based. But it isn't. DV - the once dominating format in camcorders, now less so, was probably the last format that used only frame-based compression. Some QuickTime variants too, and when you used these in Flash you could probably use a convert-to-frames method there. Modern formats use temporal compression which gives smaller files and better quality. This includes FLV. Temporal compression means - in short - that most frames are constructed from bits and pieces from previous and upcoming frames. Nowadays a videofile is a stream of data, not a chunk of frames. This means that you can't go to a specific frame either; there is no way in ActionScript you can say video.gotoandstop(112); or similar - for the same reason. Instead you can (only) seek to its nearest keyframe, video keyframe that is. Since these are (the only) full frames in the stream, they can be displayed in a frozen state easily. In an editing app you need to reconstruct the missing full frames before cutting is possible. Flash doesn't have this feature, and so you can't convert video to individual frames there. Video editing apps however will be able to export individual frames as bitmaps, easily. This batch can be imported to Flash. The whole process is pretty easy, and in some applications it may be the best way to go. But you lose the efficiency of the video compression. A video file say 2 MB in size may very likely end up as 10 MB of bitmaps, and may not play back as smooth either.
Not everything that can be counted counts, and not everything that counts can be counted.