Abstract

Capturing photos or videos of moving subjects with a handheld camera requires the capturing user to look frequently at a displayed preview of the captured scene to maintain subject framing. This can be difficult in many contexts and also disrupts the user’s ability to experience the event directly. This disclosure describes techniques to provide real-time, eyes-free haptic and/or visual feedback to assist users in framing and following a moving subject when capturing a photo or video. Users can select one or more primary subjects within the live preview. Upon selection, the target subject can be marked with a prominently visible visual indicator, such as a bounding box. Subject detection is then performed continually by using the camera hardware abstraction layer or application-level machine learning models to process frames in real time and the detected subject tracking indicator (e.g., a bounding box) is updated continually as the subject moves. The subject is kept within the frame by automatic zoom adjustments or digital cropping. When the subject approaches or crosses a field of view threshold, a haptic vibration, audio cue, and/or visual indicator is provided to the user. The user can then adjust the direction of the camera to re-center the subject without needing to look directly at the preview.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS