Some of the important features about Gesture Based Scroll (Using AI) addon are listed below. This is followed by an overview about this addon. This summary contains few words about the addon, its function and features. Next is FAQ, this section is useful to get information about how this addon works in your browser. If you couldn’t find a specific answer, please fill the bug report form in this page or the contact form in the homepage to let us know about your question. Following the FAQ, is the bug report form. It is designed to collect user feedbacks and bugs about this addon. Users can suggest improvements or feature requests through this form too. Next section is for submitting comments about this addon which uses facebook comments plugin. The last section is dedicated to reports submitted through the GitHub website. We hope the information provided in this page, could help you get all the basics you need about this addon and enable you to have a better experience with Gesture Based Scroll (Using AI).
Gesture Based Scroll (Using AI) is a lite addon that let you easily scroll the page with gestures; uses modern AI technology right in your browser.
4 Reasons to Use Gesture Based Scroll (Using AI)
Easily control page scroll with hand or face gestures.
Needs training (four steps) before using the addon.
Works based on images captured from your webcam.
Available for chrome only on all operating systems.
Frequently Asked Questions
1. What is Gesture Based Scroll (Using AI)?
Gesture Based Scroll (Using AI) is a browser addon that let you scroll a page with hand or facial gestures (without using mouse or keyboard). Before using the addon you need to train it first. Once the addon is fully trained, please click on the start button and make desired gestures in front of the camera (UP, Down and Stop gestures). Please make sure to place your face/ hand exactly in front of the camera (within the grey rectangle). As long as the camera is ON, it will look for your gestures and take action according to your gestures. If you want to stop the inferring process, please click on the stop button once. Read below FAQs to get more info on how to use this addon.
2. How can I work with this addon?
When you first install the addon, please click on the toolbar icon on the main UI. There are several buttons which are explained in the following. The first thing to do before using the addon is to train it for scrolling UP, Down and Stop actions. Please click on the - Train #1 - button. If the browser asks for the webcam access, accept the permission. Then, place your hand in front of the camera (webcam), and make gestures for UP scroll action. For example you can point your hand to the up direction or turn your face to the up direction. As long as the camera is ON, the addon captures shots from your gesture and uses that for training. The default speed for capturing images is 3 frames per seconds. You can adjust this speed based on your machine. But please note that, increasing the speed will reduces the training time but adds a very high computational burden to your machine and may cause your system to become unresponsive. At the current speed (3 frames per second), please continue training for around 30 seconds for each training step (#1, #2, #3 and finish-up). Once all the training steps are done, you can use the addon to infer your gestures.
3. How can I train this addon to infer my gestures?
As mentioned above, first place you hand/face in front of the camera. Then, click on the Train #1 button to start training - scroll UP - gesture. Play scroll UP gesture in front of the camera for around 30 seconds. Then click on the stop button next to the train button. Continue this procedure for Train #2 and Train #3. Then click on the - Finish up - button. The finish up will complete the training process after few seconds. When training is done (all 4 steps), you can use the addon to infer your gestures through the camera or webcam. To do so, please click on the Start button. The camera will be turned ON and looking for your gestures to take appropriate action. In order to stop the inferring process, please click on the stop button once. Also note that, this AI technology is very new and may have some bugs or technical issues. Therefore, please do NOT expect the addon to fully operate without any errors. If you see any error in the training stage or inferring step, please report it through the bug report form in this page.
4. When training is done, how can I use the addon?
When you are done with training, all four steps, please click on the start button. The camera will look for your gestures to take related actions. For example, holding your hand when pointing to the UP direction will be translated to the scroll UP action in the active tab (only active tab will listen for scroll actions). Similarly, pointing down with your hand will make the page to scroll down. When you are done with scrolling, please either close the UI or press on the stop button. Closing the UI will completely turn off the camera, but pressing on the stop button will stop the scrolling action and keep the camera ON.
5. Is this addon considered lite?
No, since this addon uses a noticeable amount of memory and CPU in the training and inferring stages, it might slow down your browser or PC. Therefore, please consider using this addon if you have a decent machine and latest OS. Moreover, please close the UI when you are not using the addon to release and keep browser resources for other tasks.
6. What is the technology behind the AI engine in this addon?
This addon uses the latest AI technology developed by Google and called Deeplearn. Deeplearn recently changed its name to TensorJS and is actively under development by Google. Many AI projects are done with this technology and many companies are incorporating it in to their products. Please read this page (https://js.tensorflow.org/) to get more info about this new AI technology. Please note that, currently only Chrome browser supports this technology. Moreover, as also mentioned above, there might be some bugs or instabilities in this addon and that is because this technology is very new and still under development. Please report any issues through the bug report form in this page.
7. How can I clear the training data?
If you want to clear the training data, please open the main UI and click on the clear button. A confirmation box will be shown before deleting all data. Please note that, when all data is cleared, you need to train the addon again in order to use it.
8. Can I use this addon for other actions, such as page back, forward, or etc.?
No, unfortunately this addon is only made for scroll action. In future, support for other actions such as page back, forward, reload, etc. will be added.
Your feedback goes a long way towards making this addon even better. With this Bug report form, you can submit bugs or suggest enhancements for the functionality or performance of this addon. Please note, this bug report form will be sent to the developer(s) of this addon. You will not get an immediate or personal response at the moment. However, the developer(s) may contact you through the email address provided in this form for further clarifications or requesting additional information regarding the reported bug(s).
Comments | Reviews
Please note: keep these reviews clean, and avoid the usage of derogatory or inappropriate language or terminology in references to individuals or groups based on race, national or ethnic origin, color, religion, age, gender, etc. Do not submit bugs in your reviews, use the above bug-report section instead. Also, never post any personal information in these reviews.
GitHub Reports | Fork me on GitHub