EQ Explorers

Inspiration

EQ Explorers is inspired by AI-generated visuals, particularly tools like Sora, which can transform text into high-quality videos. Witnessing how these animations bring emotions to life sparked our idea in developing a gamified platform to help children with learning disabilities understand and detect emotions. By turning abstract emotional concepts into engaging, interactive animations and scenarios, we created a web app that strengthens their emotional intelligence and makes learning both accessible and fun.

What it does

Our app leverages OpenAI technologies to help children improve their understanding of emotions and facial expressions. To simulate a live experience, we utilized AI-generated videos from tools like Sora to create realistic animations that vividly showcase various human emotions. These animations were designed to present clear, expressive facial and body movement cues that children could observe and analyze. The videos are then presented to the user in a multiple-choice format, where the child can select the emotion they believe is being expressed, encouraging them to think critically and empathize with others. Each question provides instant feedback, allowing children to learn from their responses and gradually build their emotional recognition skills.

How we built it

AI: Sora-AI: Generates scenario videos GPT-4o model: Generates text-based scenarios to feed into Sora-AI

Databases:

• Amazon DynamoDB for storing text-based data
• Amazon S3 Bucket for storing video-based files

Frontend: React.js Backend: Node.js

Challenges we ran into

When implementing EQ Explorer, our original vision was to create an adaptive system that would dynamically generate new sets of problems tailored specifically to each individual user. The plan was to incorporate picture-based questions, where each scenario would be paired with AI-generated images showcasing various emotions the character might feel afterward or vice versa. However, the performance and accuracy of DALL·E 3 via the API differed noticeably from its functionality when used within its built-in interface in ChatGPT. In the chatbot, GPT dynamically enhances user prompts by adding contextual details, which enables DALL·E to generate visually rich images. We attempted to do the same with DALL·E, but the limited prompt size restricted the level of detail we could achieve, resulting in lower-quality outputs that were unusable.

Accomplishments that we're proud of

Despite the challenges we faced with using DALL·E, we successfully adapted and pivoted to SORA AI, which delivered even better results and enhanced media to highlight the issues we aimed to address. We are also proud of combining our diverse backgrounds and expertise to build a full-stack project that effectively leveraged each of our strengths. Above all, we take the greatest pride in our platform, which stands as a testament to our collaborative effort and commitment to creating something that will truly help children adapt more effectively in society.

What we learned

We learned that careful planning is essential, particularly when it comes to technical details such as selecting the right technologies, ensuring they are accessible and cost-effective, and evaluating how well they integrate. This approach was crucial in the successful development of the project we are presenting. We also gained valuable experience in adapting to challenges that arose along the way. Furthermore, we expanded our skill set by learning new technologies, including Sora-AI and several AWS services, which we had not worked with before.

What's next for EQ Explorers

The next steps for EQ Explorers involve adding scoring and levels for individual users, which will also require creating accounts to track their progress and performance throughout the exercises. Additionally, we plan to leverage AI to generate a wide variety of new questions, ensuring users always have fresh challenges to improve their emotional recognition skills. While it won't be endless, the goal is to provide a large, diverse selection of content for users to practice and develop their abilities. Our ultimate aim is for this platform to assist children in better understanding and recognizing human emotions, helping them navigate social interactions with more confidence.

Built With

• amazon-web-services
• express.js
• openai
• python
• react
• sora