PhotoTherapy

Our project aims to provide users with a source of entertainment and relaxation to help them de-stress. According to the National Institutes of Health, individuals with depression might thus benefit from additional training in generating vivid imagery for positive events. Our project allows you to either upload photos manually or connect your google photos to our application and be able to scroll through them. A description of the atmosphere and the sounds in the image is generated using GPT 4o, which is then sent to the eleven labs api and the suno api, which create a short effects sound and a longer vibe song respectively, allowing you to experience you photos through music and perfectly complement the nostalgia of looking through photos. Additionally, our app has the functionality to edit your images and try out goofy transformations, whether that means picturing how the image would have been drawn by Van Gogh to seamlessly incorporating a picture of a cat on a flying horse to your images. We used a custom style transfer neural network with DALL-E to perform style transfers, and DALL-E would generate the style images that would be combined with your photo to create the styled output. As for the image editing functionalities, we built a pipeline to generate the photo with DALL-E, remove its background, and place the image in the user-specified location. Looking at photos can be nostalgic and bring up a lot of emotions, and for this reason we built a mental health chatbot with langflow that is connected to a vector database of mental health documents in astraDB that detail best practices and guidelines. Our entire app was designed with an easy to use and seamless UI in Gradio.