Help for visually impaired people
Revilio is an innovative application designed to improve the quality of life of people with visual impairments. Our technology helps to recognize objects and text in real time. The application processes information exclusively on the user's device and does not require an Internet connection.
Main functions
Text reading: text recognition using the device's camera and reading by a voice synthesizer.
Text Search: Search for user-specified text among all text in the camera image
Object Search: Detect a user-specified object using object recognition technology in the camera frame
Technologies
Revilio uses the built-in highly efficient Apple iOS technologies for text, object recognition and voice interaction, optimized for iPhone and iPad hardware to provide the best performance available on your device:
- Vision Framework and Optical Character Recognition (OCR): Using Vision technologies to analyze images and recognize text allows you to quickly extract the information you need
- CoreML Framework and Object Detection: Implementing CoreML to implement machine learning models that can detect and classify objects in real time, such as recognizing objects in your home environment.
- AVFAudio Framework Voice Synthesizer: Implementing AVSpeechSynthesizer technology to read out text and app notifications.
- Speech Framework and Voice Recognition: Using SFSpeechRecognizer technology to convert speech to text, providing convenient voice input and interaction with the app, as well as support for hands-free commands.
- Siri Integration: Implementing shortcuts using the AppIntents framework to quickly call app functions via Siri voice commands, as well as customizing and creating custom commands using the Shortcuts app for increased personalization and usability.
Features
1. Object search
- Press the function call button and say the object you are looking for out loud (from the list of supported objects).
- The application will recognize speech, match it with the name of a specific object and start searching for this object in the camera image in real time.
- When an object is detected, it will give tactile (vibration) and sound (synthesized speech) signals indicating the location of the object in the frame relative to the center of the camera. You can change the type of signals in the application settings menu.
- You can change the object selection method in the application settings by specifying a list instead of voice input. In this case, when you call the function, the application will display a menu with a list of all supported objects. The menu allows you to filter the list by entering part of the name of the desired object using the keyboard. Select the desired object from the list and the application will immediately begin searching for it.
- Siri integration: сall Siri, say "Revilio, find object spoon" or other supported object, and App will start search for it immediately.
2. Text search
- Press the function call button and say the word or phrase you are looking for out loud.
- The application converts speech into text and starts searching for it among all the text in the camera's field of view.
- When text is detected, it will give tactile (vibration) and sound (synthesized speech) signals indicating the location of the searched text in the frame relative to the center of the camera. You can change the type of signals in the application settings menu.
- You can change the method of entering the searched text in the application settings, specifying the keyboard instead of voice input. In this case, when calling this function, the application will display a menu with an input field and a keyboard. The menu also contains a list with the search history of the last ten words or phrases. Select one of the items from this list, and its text will appear in the input field, or enter the desired text using the keyboard. Click the "Find" button on the keyboard or in the upper left corner of the menu, and the application will start searching for text from the input field. Click the "Clear history" button above the list with the search history or in the application settings to permanently delete the entire history of searched texts.
- Siri integration: сall Siri, say "Revilio, find text". Siri will ask you to specify the text to search for. Say a word or phrase and a search will begin immediately.
3. Text reading
- Point the camera at the text and press the function call button.
- The application will recognize the text in the camera frame and start reading it out loud with synthesized speech.
- Playback control: Use the "FORWARD" and "BACK" buttons on the main screen to go forward and back through the recognized text, the transition step - a sentence or a line, can be selected in the application settings. Use the "PAUSE"/"RESUME" button to pause and continue reading. Use the "READING SPEED" button to select the speech speed of the synthesizer - normal or accelerated.
- Segmentation function: Automatically determines the main text block in the center of the frame (for example, the desired page of a book in a spread), ignoring neighboring ones, to focus on the relevant content. The function can be disabled in the application settings.
- Siri integration: сall Siri, say "Revilio, reading", and the reading function will start immediately.
Additional application features
- Complete autonomy: all image, video, audio (speech recognition) and data processing occurs exclusively on your device. Data is not saved to disk (except for the local history of Text Search and application settings) and is not transmitted to the network.
- Select a sound source: quiet or loud mode. In order to direct the application sounds to the earphone or headset connected to the device, select the "Quiet" mode.
- Pause and resume: During any of the three main functions, press the "Pause" button to pause the operation. Press the button again to immediately resume the function. To save the device's battery and protect against overheating, it is recommended to pause or stop the Object Search and Text Search functions whenever they are not currently in use.
- Intuitive control via voice commands and haptic feedback.
- Control the camera zoom using the buttons on the main screen of the application for better recognition of objects or text.
- Use the flashlight (flash) during the main functions to get better results in low light conditions, you can activate this feature in the application settings menu - separately for each of the three main functions.
- Use built-in Apple technologies (Core ML, Vision Framework, Speech, AVFoundation) to provide fast, private and accessible help in everyday tasks.