SuRe VISION

Most Home appliances nowadays are simple implementations of complex Industrial solutions. Security and Surveillance is one such domain which evolved from industry usage to common home safety usage. Home security solutions today come with hefty price points for devices and subscriptions towards data management and backups, as huge data is generated and need to be stored. We tried to find the solution of this problem by harnessing the technological capabilities and implemented a deep learning-based solution and bringing down the cost of both hardware and storage. For high safety, the HD cameras are used everywhere and the video files size is big due to the high-resolution data. Currently surveillance camera manufacturers use compression techniques to reduce the data size and bring down the file size to an optimum level where they can be viewable, but still the data size is large. With Deep Learning based Super Resolution algorithms like EDSR, WDSR, SR-GAN we can compute and generate the high quality videos from low quality videos. Our approach is to store the data with low resolution in less space and generate the high-resolution videos on demand (only when we suspect a security breach or an anonymous activity) this reduces the storage cost a lot as these kinds of situations will raise only once in a while. We also tried to implement a low- resolution camera which will serve the purpose without compromising the security.