A Collection of Pre Trained Models
It provides pre trained models so that you can customize further with your data.
Radiology Use cases
This use case is to determine the severity of various deceases such pneumonia, cystogram, fluoroscopy,, etc. It uses the pre trained deep learning models and intuitive ui for the doctors which enables the doctors/radiologists to detect the finer details of the X-Ray images with high accuracy,
Biopsy Image Annotations
This use case is about using Deep learning models on high resolution MRI/Biopsy images. Detecting various glands and cell types and classify them based on their patterns. Calculate the area and concentration per micro meter level measurements. which correlates various cancer types and severity of the cancer.
This can be used for breast, lung, prostate , thyroid and throat cancers and many more. There is quite lot academic research has been done and NIH provides both guidance and data sel. Hospitals and Pharma have large data set as well.
There are open source efforts to model this pattern. where as AI-OBJECTS done the models and pre trained. It can be further trained and customized to fit the need.
UI/UX allows you to zoom and panning, creation of new annotations, comparisons, deletions and updates. Ability to track the progress of the patterns and measure accurate concentrations using map view.
Deep fracture analysis
This use case is for orthopedic doctors. Deep learning models are used on the 3D X-Ray images and measure the micro fractures and perform deep analysis find out the nerve pains caused by small deformity,
Deep Skin Analysis
This use case is for Dermatology department . Deep learning models are used on images taken by the user. User interface provides the ability focus and get high quality pictures. Once high quality pictures are taken is been sent to the AI inference engine for analytics. The analytics report contains the various data point including the scoring, vitamin deficiencies. Statistical comparison to the normalcy will help doctors to treat the symptoms. Skin care product companies can leverage this capability to provide precision product to the consumer.
Deep Blood Loss Analysis
This use case is for Surgeons . Deep learning models are used on the images taken at the blood wet towels at the surgical table. System automatically weighs the towels and performs inferences the images and measure the accurate blood loss.
Automatic Speech Recognition
Developed custom model on top of the DeepSpeech models and improved accuracy and decreased the word error rate to 12%. It is developed to improve accuracy on STEM related subjects such as Biology. It better performed compared to Google and IBM Watson on our Biology test data set
Chatbot for your Services
Developed custom NLP and NLU based models on top of BERT algorithm.
Data Analytics & Predictions
Custom developed and fine tuned models for the fraud detection, email classification, ticket classification, review classification and predictions