Our Prize Winning AWS Hackathon Entry

This text is written to explain my project that I submitted to the AWS Hackathon on Devpost. Me and my buddy had preliminary started the undertaking as a learning step to build and deploy a machine learning challenge. We ultimately ended up winning the runner up for Computer Vision for our efforts. With fast urbanization and industrialization Air pollution has change into an alarming challenge globally. Among completely different air pollutants particles , matter in (PM2.5) size vary are ready traverse deeply into our respiratory tract , transmitting hazardous chemicals into the human lungs and BloodVitals SPO2 blood inflicting respiratory and cardiovascular health points . We goal to develop image-based air quality analysis software , in particular a one which estimates the focus of particulate matter with diameters in range of 2.5 micrometers. It will possibly tell you what your respiration from photos generated using shopper grade hardware. Convolutional Neural Network (CNN) with the excessive computational power of Amazon EC2 DL1 occasion to unravel a computer imaginative and prescient downside of classifying natural photos into three completely different categories ;healthy , reasonable and hazard based mostly on their PM2.5 concentrations.



We created our personal dataset by combining data from 2 sources . We learned that, wireless blood oxygen check some reasons for overfitting could possibly be high imbalance in information , we had 2232 wholesome images the place as the average and hazard classes had only 1480 and 591 images respectively . 2232. Then we trained our model following the identical technique as earlier than however this time we used ImageNet weights un-freezed the final 100 layers for coaching , then high-quality-tuned the mannequin with a lower studying rate ,we additionally used early stopping callback with a patience of 3 . Now, there are some extra things we have now performed to improve our mannequin. We migrated our model onto the dl1 instance and initially trained using 1 gaudi processor. This considerably improved our efficiency and effectivity .We were in a position to rapidly check , BloodVitals SPO2 experiment and tweak modifications . 76 %. We then transformed our mannequin into a tensorflow lite mannequin and used a huggingface demo tool called gradio to rapidly display our utility.



Initially as we tried coaching our model we observed among the picture information within the dataset had been corrupted .We automated the means of scanning by and deleting these recordsdata and then proceeded to training mannequin . We used EfficientNet architecture since its renown for its performance in low spec machines, this is known as switch learning . 30 layers and set the weights parameter to false. 60 % but our training accuracy was excessive, the model may very well be overfitting. As we've got seen we were able an accuracy of about 76 % which is pretty good considering the data we had. Now , BloodVitals SPO2 why do we want this software and how is it completely different from the existing air monitoring system systems, we already know Exposure to advantageous particles could cause long term health results reminiscent of lung and BloodVitals SPO2 coronary heart problem however they also trigger quick time period well being results corresponding to eye, nostril, throat irritation.



The present air high quality monitoring methods are highly dependent on monitoring stations, that are again situated far away because of the excessive setup value and BloodVitals SPO2 expensive gear. For a particular location, the air quality is determined by how air strikes by the area and the way people are influencing the air , so we can not depend on one thing that was predicted some time back or BloodVitals SPO2 in some place that's not our site of curiosity .This software helps you analyse the air quality round you with a single image and extra importantly is portable. We are able to carry it wherever we wish to and know what your being uncovered to . Honestly I don’t know, I think there's potential for this app to combine into our routine. People may want options to help plan their day out, integrate with their phone’s calender and recommend the most effective time to plan an outdoor exercise. Right now the Artificial intelligence that runs the machine runs on the azure cloud. Sooner or later we would like to have the ability to carry it into the app and run it natively.



Certain constituents in the blood have an effect on the absorption of light at numerous wavelengths by the blood. Oxyhemoglobin absorbs mild more strongly within the infrared region than in the purple area, whereas hemoglobin exhibits the reverse conduct. Therefore, highly oxygenated blood with a excessive focus of oxyhemoglobin and a low focus of hemoglobin will tend to have a high ratio of optical transmissivity in the crimson region to optical transmissivity within the infrared area. These alternating parts are amplified and then segregated by sampling gadgets operating in synchronism with the crimson/infrared switching, BloodVitals SPO2 so as to provide separate alerts on separate channels representing the red and real-time SPO2 tracking infrared gentle transmission of the body construction. After low-cross filtering to take away sign components at or BloodVitals SPO2 above the switching frequency, each of the separate alerts represents a plot of optical transmissivity of the body construction at a selected wavelength versus time. AC component prompted only by optical absorption by the blood and various on the pulse frequency or BloodVitals SPO2 heart price of the organism.