Defining and detecting harassment through cameras is technical challenge

[viththagi]

Defining harassment through objective criteria for video analytics is complex. Subtle cues or non-physical harassment might be difficult to detect using cameras alone. Non-verbal Cues: Often, harassment is communicated through non-verbal cues like body language, facial expressions, or tone of voice. These subtleties are very challenging for video analytics software to interpret reliably.

so for activity recognition:Develop algorithms that go beyond simple object detection to recognize specific actions and interactions that might indicate harassment. This includes analyzing posture, gestures, and relative positioning of people.

Sensor Fusion: Combining data from cameras with other sensors like microphones (with privacy safeguards) or wearables (with user consent) can provide a more comprehensive picture of the situation.

Advancements in AI, particularly in areas like natural language processing and sentiment analysis, can help interpret non-verbal cues and the emotional context of interactions.

so i would like to work on these sensor fusion and sentiment analysis so assign me this issue

[Neh2005]

I would also like to work on this issue. Please assign this to me under Gssoc'24.

[shashbh18]

I would love to work on it.Please assign this issue to me.