Chest-rAi - AI Based System to Assist Radiologist

Chest-rAi - AI Based System to Assist Radiologist

Chest-rAi analyses chest radiographs using machine learning techniques to identify and highlight various pulmonary and other thoracic diseases. Chest-rAi is a deep learning system blended with a traditional radiologist approach of systematically examining a chest radiograph. It is built on hierarchical classifiers compared to state of art large multi-label classifiers along with visual indicators that can highlight the regions of interest specifically. The first level will separate the normal lungs from abnormal lungs using Chest X Ray. The further levels separate the infected lungs into classes of similar looking abnormalities. Similarly, as we go down the tree, our framework tends to segregate different kinds of abnormality at different levels. At each level, the decision making is automated using a trained neural network model. It is an adjunct tool and is not intended to replace a clinician's review of the radiograph or his or her clinical judgment. Chest-rAI model (FP-32) size reduces ~50% of original size and inference time reduces by 46% of original time. Majority of the gains were observed in Densenet 121 and Densenet169 family architecture. The optimized inference pipeline for Chest-rAi was 1.84 times faster on Intel(R) Xeon(R) Platinum 8380 CPU @ 2.30GHz (code named Ice lake).

*Please note that member solutions are often customizable to meet the needs of individual enterprise end users.

CONTACT COMPANY

SOLUTION FEATURES

  • Time Saver for Radiologist
    The solution assistant improves the efficiency of the Radiologists enabling them to use their time optimally. OpenVino optimizations help radiologists by providing reference localization of findings.
  • Train Radiologist
    Chest-rAI analyses the X-ray and predicts the classification, localization and contextual report for abnormalities. With a defined input data set, Chest-rAI can be used to train the radiologists on individual or combination symptom detection.
  • Standardize Report writing
    Chest-rAI is also expected to standardize report writing for chest X ray diagnosis.
  • Report from Anywhere
    Chest-rAI works on cloud-based technology, and the presence of a radiologist in the Lab for reporting is not mandatory. Using Chest-rAI radiologists can report the X rays from any location with the help of internet support. This is encouraging WFX.
  • Benefits to Patient
    It facilitates better care by fast triaging and providing more accurate radiological findings. Also with the help of OpenVINO, patients were able to get reports faster

CATEGORIES

Combatting Covid-19IndiaAI Software/SaaSHealthcareMedical imaging, analysis and diagnosticsCSP - Microsoft AzureIntel® Xeon® Scalable ProcessorIntel® Distribution of OpenVINO™ Toolkit powered by oneAPIModels can be trained - requires labeled dataLinux LSTMOtherProprietaryDeep Learning