The problem Emotion Powered AI Chatbot solves

The Problem at Hand:
Conventional chatbots often lack the ability to understand and respond to the emotional nuances that are integral to human communication. Users often find themselves engaging with sterile and robotic interactions, missing the depth and authenticity that emotions bring to a conversation. Moreover, in the absence of emotional context, these chatbots may fail to provide relevant and empathetic responses, leading to a suboptimal user experience.

The Emotion Powered AI Chatbot addresses these challenges by introducing a dynamic layer of emotional awareness to digital conversations. By leveraging facial emotion detection, it gauges the user's emotional state in real-time, creating an opportunity for more meaningful and personalized interactions.

Enhancing User Experience

1. Personalized Responses:
   The chatbot tailors its responses based on the user's detected emotions. For instance, if the user expresses happiness, the chatbot adapts its tone to be more cheerful and uplifting. Conversely, if the user appears sad, the chatbot might offer comforting or empathetic responses. This personalization significantly improves user engagement and satisfaction.

2. Effective Communication:
   Understanding emotions is crucial for effective communication. The chatbot's ability to detect emotions ensures that responses are contextually appropriate, fostering a more natural and human-like conversation. This is particularly beneficial in scenarios where conveying the right emotional tone is essential, such as customer support or mental health chat applications.

3. Emotionally Intelligent Assistance:
   In tasks requiring emotional support or guidance, such as mental health discussions, the Emotion Powered AI Chatbot becomes a valuable companion. It can adapt its responses based on the user's emotional needs, providing a supportive environment for individuals seeking understanding or assistance.

Challenges we ran into

1. The most difficult part was training the emotion detection model as we chose a stable dataset from the year 2013. Training it on our local machines was a mess. So, we used pretrained models for gender and age detection.

2. It was hard to find a low latency api based llm for our chatbot system but at last we used the Cloudflare Worker AI API key to get Llama2 model as our chatbot.

3. Clash between Streamlit and OpenCV libraries. The FPS of the video capture decreased due to that purpose.