The popularity of computer vision and deep learning based applications has increased and used for many areas where human privacy is also concerned. In today's world, the data is very important for detailed analysis as well as retraining a model to improve classification and detection accuracy,

Hence the identity of the person can be protected by blurring their faces or replacing the face with artificially generated face. The face swapping technique can be very useful when any user uploads any private videos.

The doctors in Australian hospitals uses mobile application where they can capture the document using mobile application and key words are extracted from the document using some intelligence. The challenging part of the project was to detect the document boarder accurately and apply perspective transformation to correct the shape of the document.

Once document is corrected, it's uploaded to AWS S3 bucket where OCR is applied and send back to mobile application fill the fields with respective key wrods like name, surname, sex, and other details.

Image-based localization is a problem of predicting location and orientation of a device (i.e robots) using image captured by that device. One of the most popular application of such system is Visual Positioning System.

If you are lost in some strange city then you can capture an image of your surrounding and VPS will identify your location with orientation (as it is required to provide navigation functionality).

MapNet and its variants are state-of-the-art models in this field. MapNet is built on top of PoseNet and its variants. PoseNet and its variants are the early methods in this field which uses deep learning based approach.