Mindly

Inspiration

The inspiration behind Mindly stems from the growing prevalence of autism, which now affects approximately 1 in 6 children in the United States. This alarming statistic highlights the pressing need for improved treatment approaches that can address the unique challenges faced by individuals with autism spectrum disorder (ASD) and support their overall well-being. Patients face challenges when self-reporting symptoms to doctors, and current methods of treatment for autism often overlook the visual elements that play a crucial role in understanding and addressing the unique needs of individuals with ASD. As a result, patients are often times provided inaccurate treatments that often times worsen their conditions.

What it does

With Mindly, through videos uploaded daily, doctors can now have a comprehensive view into a patient's behavior by analyzing these videos, which leads to an enhanced treatment approach. Mindly provides comprehensive treatment because it helps clinicians track patient symptoms even outside of the clinic. Video data that is collected consistently over time reduces self reporting by patients, which prevents the human error and biases that occur in traditional assessments. Our AI algorithm analyzes uploaded videos to identify common physical symptoms related to ASD, like repetitive hand flapping. It then compiles the videos into a condensed clip, highlighting the specific moments when the symptom is present, and then rates and tracks symptom severity over time. Mindly's AI-driven capabilities revolutionize assessment, tracking, and treatment of ASD, enabling patient care to be treated with higher efficiency and accuracy. This visual data provides a valuable window into the daily experiences and challenges faced by individuals with autism, allowing clinicians to tailor treatments with greater precision and effectiveness.

How we built it

Mindly incorporates pose detection technology to effectively identify and track physical symptoms associated with autism spectrum disorder (ASD). This innovative feature enables the platform to detect and analyze specific body movements, postures, and gestures that are indicative of ASD-related behaviors. We analyzed key points and joint angles in the body, capturing the unique postures and movements of individuals with ASD. We then processed this data by Mindly's AI algorithms, which can detect patterns and anomalies in the physical behavior of patients, as well as the severity of that specific symptom. By continuously monitoring these physical symptoms over time, clinicians gain valuable insights into the effectiveness of treatments and interventions. By leveraging pose detection algorithms, Mindly can detect physical symptoms such as repetitive hand flapping, rocking, or other repetitive movements that are commonly observed in individuals with autism. Mindly's application of AI provides clinicians with objective and quantifiable data, allowing for a more accurate assessment of symptom severity and progression.

Challenges we ran into

Individuals with autism exhibit a wide range of body postures and movements, making it challenging to create a robust algorithm that can accurately detect and track these variations. The algorithm needed to account for the diverse range of poses seen in different individuals with ASD, including those with more subtle or atypical movements. To provide a seamless user experience, the algorithm needed to perform pose detection in real-time. This required optimizing the algorithm's efficiency and ensuring it could handle the computational demands of analyzing video data in near real-time without sacrificing accuracy. Video data often contains noise, such as lighting variations, occlusions, or cluttered backgrounds, which can hinder accurate pose detection. The algorithm had to be designed to handle these challenges and robustly extract relevant pose information despite potential disruptions in the video footage.

Accomplishments that we're proud of

We are proud to say that this model accurately detects the severity of one of the most common ASD symptom: repetitive hand flapping. Additionally, the severity of the symptom can also be detected from the video, and can be seen through visualization metrics displayed on Mindly's platform.

What we learned

Building Mindly required a deep understanding of the complexities of autism spectrum disorder (ASD). Our team gained a greater appreciation for the wide range of symptoms, behaviors, and challenges faced by individuals with autism. This understanding helped shape the development of the platform to address the specific needs of this diverse population. Additionally, developing Mindly provided an opportunity to explore and leverage advancements in AI and computer vision technologies, particularly in the context of pose detection and symptom analysis. Our team gained insights into the capabilities and limitations of these technologies and explored ways to optimize their performance for accurate symptom evaluation.

What's next for Mindly

The main focus for the next 6 to 12 months will be on beta testing and piloting to ensure the app is of high quality and meets the needs of users, as well as initially launch our product to clinicians all over the United States. The goal is to build a sustainable product that can be integrated into existing online healthcare systems. The focus on beta testing and piloting is crucial in ensuring the app meets the desired standards. This stage will provide valuable feedback from users, which will be used to improve the app and make it more user-friendly. Additionally, piloting the app in a real-world environment will help to identify any bugs or issues that need to be addressed before the official launch. In parallel to the development and testing process, we also need to concentrate on the administrative and marketing aspects of the app. This will involve reaching out to potential users and partners to promote the app and increase its visibility. Marketing efforts will also help to create brand awareness and build a strong user base. We hope to eventually partner with existing EMR companies, as well as provide anonymized patient data to ASD and other neurological disorder research companies.

Built With

• pose-detection
• python
• tensorflow