Introduction
VIDIZMO possesses the capability to intelligently identify activities through the utilization of an advanced video classification model. This activity recognition involves a systematic analysis of the complete video, processing 16 frames in each iteration to meticulously identify and categorize activities specified by the user. Notable activities available for detection include robbery, shopping, and trespassing.to identify and categorize activities specified by the user meticulously
For a comprehensive understanding of the Activity Recognition concept, we recommend referring to our detailed article "An In-Depth Understanding of Activity Recognition" for further insights.
Prerequisites
- Users with a Moderator+ role can generate AI Insights in the Portal.
- Users must have the Manager+ role to enable and configure Activity Recognition in VIDIZMO, ensuring the activation and proper configuration of the VIDIZMO Indexer App. For detailed instructions on configuring the VIDIZMO Indexer App, please refer to the document titled "Configuring VIDIZMO Indexer for Object Detection."
Generating Activity Insights with Automatic Processing
- Enable automatic processing in the VIDIZMO Indexer app to facilitate the system's automatic recognition of activities such as robbery, trespassing, and shopping when uploading media files.
Note: Activity Recognition functionality is exclusively applicable to video content within VIDIZMO.
2. For users on a DEMS Portal, initiate the video upload process by selecting New from the header menu and utilizing the Upload Evidence option. EVCM Portal users can achieve the same by using the Add Media button.
3. Upon processing your uploaded media file, navigate and access the Insights tab on the playback page to view the generated Activity Insights.
Refer to our documentation titled "How to enable video insights" for guidance on enabling video insights.
4. Click on Activity Insights to navigate to the timestamp of that particular activity in your video.
Generating Activity Insights with On-Demand Processing
With on-demand processing, you can selectively choose AI insights to generate on your files. This means that on-demand activity insights can be generated or regenerated on your videos. VIDIZMO offers two ways to do on-demand processing: the first is via the Process Modal, and the second is via the Custom Settings Upload.
On-demand processing using Process Modal
- Navigate to any video file within your Portal's Library and access the on-demand processing options by selecting "Process" from its overflow menu. Alternatively, you can process one or multiple files simultaneously by selecting them and then choosing "Process" from the header menu.
2. Within the Process Modal, choose the specific Activities for which you intend to generate AI insights.
3. Initiate the on-demand processing for your video by clicking the "Start" button.
4. The AI insights can be viewed on the playback page, as detailed earlier in this document.
On-demand processing using Custom Settings Upload
An alternative method for on-demand processing involves the use of Custom Settings Upload. This approach allows users to choose the AI insights generated on files during uploading. Users holding the Manager+ role can activate this option through the Upload Settings located under the Library section within your Portal Settings.
If you need a detailed understanding of uploading custom settings, please refer to "Understanding Custom Upload Media."
1. Upon enabling custom upload, DEMS users can utilize "Custom Settings Upload" (replacing 'Upload Evidence) by selecting "New" from the header menu. EVCM users, on the other hand, will observe the addition of a Process tab when uploading through the Add Media button.
2. In the Process tab, choose the specific activities for which you wish to generate AI insights.
3. Click 'Save and Process' to begin the on-demand processing workflow for your video.
