Video captured by security cameras is of low resolution and quality, especially in low-light environments. Currently, such cameras do not have the capability to enhance the recorded videos to fix such issues. While cloud-based video enhancement is feasible it is computationally resource intensive and slow, making it unsuitable for use in enhancing security camera footage. This disclosure describes techniques to selectively apply video enhancement to timeboxed segments within the video content recorded by a security camera. A human operator or a computer-vision based automated technique identifies a segment or a set of segments related to an event of interest within recorded video. The raw captured video corresponding to the timeboxed segment is provided to a cloud-based video enhancement model. The model outputs an enhanced video with better visual detail. Selectively enhancing segments of interest optimizes the use of computational resources and network bandwidth.

Creative Commons License

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