Page 314 - AI Ver 3.0 class 10_Flipbook
P. 314

The app uses a technology called Optical Character Recognition (OCR), which allows it to "see" the text in the
              image. OCR scans the letters and words in the picture, recognising them just like a human would read them. So,
              when you point your camera at a sign or document, the app identifies the words automatically. It then applies
              Augmented Reality (AR) to seamlessly overlay the translation onto the live image, so you can read the translated
              text in real-time. The combination of OCR and AR enables a highly accurate and smooth translation experience,
              making it convenient for travellers, students, and anyone in need of instant language translation.


                            Brainy Fact



                    Google Translate  was  first  released  in April  2006  and  originally  used  statistical  machine  translation  (SMT)
                    before switching to AI-powered neural translation (NMT) in 2016 for better accuracy.




                              Reboot


                     1.  How do we see and understand objects?


                     2.  Do you think the quality of the camera used in Computer Vision systems matters?






                       Computer Vision Tasks

              The applications of Computer Vision rely on a set of tasks designed to extract specific information from input
              images. This information can either be used directly for predictions or serve as a foundation for further analysis.
              These tasks play a critical role in enabling machines to interpret and understand visual data. By processing and
              analysing images, Computer Vision systems can perform a wide range of functions across various domains. The
              tasks involved in Computer Vision are:


                                                     Computer Vision Tasks



                                For Single Objects                            For Multiple Objects


                                        Classification                               Object Detection

                                      Classification and                           Instant Segmentation
                                         Localisation
              Single Objects


              This means giving one image as input to the Computer Vision application. It can be further divided into two
              categories i.e, classification  and classification + localisation.





                    312     Touchpad Artificial Intelligence (Ver. 3.0)-X
   309   310   311   312   313   314   315   316   317   318   319