Offline experience
When on-site, whether in remote areas or in underground facilities, the field engineers often find themselves without a network connexion. Deepomatic Lens has been designed to let them work without discontinuity even when they are offline.
Multiple features support this goal :
The offline mode when the connexion is not good enough
The pin work-order feature to always have the key work-orders available even offline
The offline work-order experience without on-device AI
The offline work-order experience with on-device AI
Activation
Experience
Offline mode
When the connexion is not good enough, the app changes its behaviour to improve the experience even without network.
A banner lets the user know that the app is offline
The app stops to exchange data with the server
The analysis cannot be executed online, if the on-device AI is activated, the analysis are executed on device, otherwise, the pictures are saved and the analysis will be executed when the user is back-online.
When the connexion is close to the offline threshold, the field engineer may be disturbed by offline/online mode interuption.
To avoid this, the user can lock the offline mode for a certain period of time. By clicking on the offline mode banner, he opens a modal which let him choose the duration for which one he wants to lock the offline mode. When the offline mode is locked, he can unlock it and try to reconnect.
The offline mode can be locked for up to 3 hours, after this delay, the app will try to reconnect and synchronized the pictures and AI results.
The on-device analysis can only be activated on supported device. See technical requirement
Pin work-orders
When opened, the work-orders stay in the cache until the cache is cleared. To ensure that a work order is accessible while offline, the user can pin the work-orders.
To pin a work order, from the work-orders list, the user has to slide it on the left and click on the pin work-order button. All pinned work-orders are marked with a red pin and can be easily searched by click on the pinned work-order filter.
Pined work-orders are always up to date and accessible on the user device.
Offline mode analysis
When the app is offline, the communication with the server are stopped. If on device analysis are activated (enabled for the organization and activated on the device) the offline experience is very similar to the online experience.
A. Without on-device analysis
When the user is offline without on-device analysis, he can still submit a picture but the analysis will take place when the user is back online. Hence, the results of the picture will not be available instantly. On the home page, the user can see the number of analysis to be synchronized.
When the user is back-online, the analysis synchronization is processed and the user receives notifications for all negative analysis results.
B. On-device analysis
On-device analysis let the field engineers have a similar experience when they are online and offline. When they are offline, the AI runs on device and to provide them analysis results when they need it, on the field, doing their work.
On-device analysis must be enabled for your use-case and acticated on their phone by each field engineers who wants to use it.
Indeed, to execute AI on device, the Deepomatic Lens app must download computer vision models and AI scripts dedicated to their use-case. Even if it is enabled for the use-case, if the field engineer does not activate it, the analysis will be run online when the user finds a connection.
The on-device analysis can only be activated on supported device. See technical requirement
1. Activate on-device analysis
To enable the on-device AI, the Deepomatic team will update the computer vision configuration. Once updated for a certain use-case the field engineers will see an activation button either when they sign in, either on the home page.
When the user activates the on-device AI, the Lens app downloads the AI models and script than will be executed on device. This offline package can weight up to 500 MB as there can be multiple computer vision models.
The download and activation can take a few minutes and it is key that the user stays on the download page untill the activation is complete.
2. On-device analysis experience
Once the on-device AI is activated, the user can executed offline analysis as if he were online. The user can update and modify the AI results. The results of the AI and his potential editions will be synchronized when he comes back online.
Last updated
Was this helpful?