Skip to main content

Purpose of CCTV CV Streamlit Apps

AI-prediction.py​

The AI-prediction app is used for performing inference using a selected model on an frame by frame basis. There are options to select the model to be used for performing inference, the utility, the dataset corresponding to the selected utility, and the video_group corresponding to the selected dataset.

There is also a filter which allow the user to only look at the frames with inspection defects.

Defect Type Classifier (based on defect frames) / Defect Type CLassifier (based on Videos)​

The Defect Type Classifier apps are used to classify videos into different `video_type's based on a frame from the video or by watching the entire video itself. The video_type of a video is used in the CCTV CV processing pipeline and is important for performing Azure_OCR. Currently takes in CSV files to figure out the paths to the frames/videos

Distance Region Extractor and Video Size Determiner​

The Distance Region Extractor and Video Size determiner App is used to store the location of the distance annotation for a given frame. It is also used to determine the dimensions of the video.