Apps used to recognize objects and/or read text for peoplewho are blind and visually impaired have increased in number. We have discussedTapTapSee recently, an others exist as well.
This post details iDentifi, a new free app that attempts todescribe objects and read text for people who are blind and visually impaired.
What is iDentifi?
Anmol Tuckrel, a high school student from Toronto, Canada,began work on the app about a year ago. According to a TechCrunch article, Tuckrelwas fascinated by the possibilities of machine learning and computer vision.The app uses Google Vision, CloudSight and Google Translate, all trustedresources that can distinguish objects easily. These facts indicate thatiDentifi uses artificial intelligence to identify objects whereas apps likeTapTapSee use crowdsourcing.
Using the App
Before attempting to use the app, please note that you must be connected to the internet to use it. The app’s layout is quite easy to comprehend. Its initialscreen contains four buttons, one in each corner of the screen–“Settings” inthe top left, “Instructions” in the top right, “Select photo” in the bottomleft and “Take photo” in the bottom right. Of course, if you flick left andright, you will locate the same buttons in the same order. Knowing theirlocation, however, allows you to find the button you want without extra flicksor swipes.
Each button and the area surrounding it is brightly coloredwith a different color included for each button or area of the screen. As aresult, people with low vision can distinguish the buttons easily, andindividuals who use both VoiceOver and their remaining vision benefit since theapp’s functionality is excellent in both cases.
If you press the “Settings” button, you first choose thelanguage for all interactions with the app from the list of over 25 languages.Next is the mode button where you choose from “Images low accuracy”, “Imageshigh accuracy”, or text. The low accuracy mode provides a general descriptionof the picture you take and returns the quickest response. The high accuracymode gives you a more detailed description of the image and requires more timefor receiving a response. In text mode, the app tries to read the text from theimage you’ve taken.
The final setting is speaking rate—how fast you want the appto speak to you when it reads its results; the settings are very slow, slow,normal, fast, and very fast.
The instructions describe some of the app’s functionalityand tell you the location of important buttons on the app. The instructions donot stay on the screen, but if you need to hear them again, double tap theinstructions button a second time.
Selecting a photo sends that photo to the app; iDentifi thentries to determine what is in the photo. You must allow iDentifi to access yourphotos and also the camera. Once you hit the select photo button, you see thestandard camera interface that you would use to send a photo to Facebook,include one in a message, etc.
When you double tap this button, you see a screen thatmimics the standard iPhone camera screen with buttons for flash, viewfinder, cameramode, camera chooser, take picture, and cancel. If you are satisfied with thecamera settings, double tap the “Take picture” button, located just above theiPhone’s home button. You will hear a sound as the phone takes the picture. Youthen can select “Retake” or “Use photo”, found on the bottom left and bottomright of the screen respectively. If you have usable vision and believe that yourpicture is not satisfactory or if you just want to use a different picture,select the retake button and start the process over.
If you tap on use photo, you hear the app say, “Loading”. Atthis point, the picture runs through the app for identification purposes. Youcan retake a picture as many times as you like, but you must hit the use photobutton for the app to begin the identification process. All photos you takeusing this app are not saved. The identifications given by the app are not ableto be reread and do not remain on the screen, but you can try the three-fingerquadruple tap gesture to put the response on the clipboard and add it to amessage, email, etc.
Currently iDentifi is available on the iOS platform only;the developer plans to create an Android version in the future. Because theresults of the picture recognitions are not shown on the screen, individualswho are deafblind and anyone using a braille display may have problemsaccessing the results. The app will read text and does a good job doing so. Itwill not replace an OCR app like KNFB Reader, though, especially if you storefiles for later reading. If you don’t need to store the file or go back andread it multiple times, iDentifi will work well.
The developer hopes to increase the available languages toclose to 100 and wants the app to work in video mode. He appears to beresponsive and open to suggestions so send them and help improve the app.
Finally, the only other limitation, as is the case with allcamera apps, is the ability of each person to take a suitable picture. Fortunately,you do not have to have the camera perfectly centered to take a usable picture.
The iDentifi app is an excellent choice for anyone who isblind and visually impaired. It identifies objects quite well and reads textreasonably well also. Remember to turn the mode to text if you want the app toread text; otherwise, it will simply tell you there is text without reading it.The results with the mode set to high accuracy are very good; its descriptionsof objects and their colors are quite helpful. You may find that this app alsodoubles as a color identifier, at least for basic colors. Would you like to seethe app in action? Watch thisshort video and view another one included in the TechCruncharticle written about the app. For more information about iDentifi, visitthe website at http://getidentifi.com/#home-section.The site discusses the numerous awards and the press coverage the app hasreceived and tells you how to get support or make comments about the app. Getthe app at the following link or search for iDentifi on the app store;remember that the d is the only capitalized letter.