Detecting Acute Lymphoblastic Leukemia Lymphoblasts with Tensorflow/oneAPI/OpenVINO & Neural Compute Stick
- 0 Collaborators
An open-source classifier programmed using the Intel® Distribution for Python* and trained using Intel® Optimization for TensorFlow*. The model is deployed on a Raspberry 4 using Intel® Distribution of OpenVINO™ Toolkit and inference is carried out using Neural Compute Stick 2. ...learn more
Project status: Published/In Market
oneAPI, Internet of Things, Artificial Intelligence
Groups
Hacker Lab IoT,
DeepLearning,
Artificial Intelligence Europe,
Movidius™ Neural Compute Group,
Internet of Things
Intel Technologies
oneAPI,
Intel Opt ML/DL Framework,
Intel Python,
OpenVINO,
Movidius NCS
Overview / Usage
IntroductionThe Acute Lymphoblastic Leukemia (ALL) oneAPI Classifier is an open-source classifier programmed using the Intel® Distribution for Python* and trained using Intel® Optimization for TensorFlow*. The model is deployed on a Raspberry 4 using Intel® Distribution of OpenVINO™ Toolkit and inference is carried out using the Intel® Neural Compute Stick 2 (Intel® NCS2).
MotivationThe motivation for this project came from the interest in exploring how Intel technologies could be used to create an improved version of the Acute Lymphoblastic Leukemia Tensorflow 2020 project. The goal is to create an improved computer vision model that was capable of detecting Acute Lymphoblastic Leukemia in unseen images of periphial blood samples with high accuracy and efficiency, running a low powered/low resources Raspberry Pi 4.
Acute Lymphoblastic LeukemiaAcute Lymphoblastic Leukemia (ALL), also known as Acute Lymphocytic Leukemia, is a cancer that affects the Lymphoid blood cell lineage. Unlike AML, ALL only affects the white blood cells, namely, Lymphocytes. Lymphocytes include B Cells, T Cells and NK (Natural Killer) cells. ALL is caused by Lymphoid Blasts, or Lymphoblasts, developing into immature Lymphocytes, and an abnormal amount of these immature Lymphocytes are produced. Lymphocytes are white blood cells and play a very important role in the immune system helping to fight off diseases. Acute Lymphoblastic Leukemia is most commonly found in children, and is the most common form of child cancer, with around 3000 cases a year in the US alone. Like Acute Myeloid Leukemia, although common, it is still quite rare. In both children and adults, early detection is critical. Treatment must start immediately due to the aggressiveness of the cancer. More info.
ALL-IDBYou need to be granted access to use the Acute Lymphoblastic Leukemia Image Database for Image Processing dataset. You can find the application form and information about getting access to the dataset on this page as well as information on how to contribute back to the project here. If you are not able to obtain a copy of the dataset please feel free to try this tutorial on your own dataset, we would be very happy to find additional AML & ALL datasets.
GETTING STARTEDReady to get started ? Head over to the official documentation for instructions on how to download/install and setup the Acute Lymphoblastic Leukemia oneAPI Classifier 2021.
Methodology / Approach
In this project we utilize Intel Distribution of Python and Intel Optimized Tensorflow to use the full power of Intel architecture and yield high performance for training. We use OpenVINO Model Optimizer to convert the model to an Intermediate Representation, allowing the model to be deployed on an UP2 and Raspberry Pi 4.
The classifier is IoT connected using the HIAS iotJumpWay broker and be can be used for inference by a number of our Acute Lymphoblastic Detection Systems.
Technologies Used
Intel® Distribution for PythonIntel® Distribution for Python enhances standard Python and helps to speed up popular AI packages such as Numpy, SciPy and Scikit-Learn.
Intel® Optimization for TensorFlowIntel® Optimization for TensorFlow optimizes the popular Tensorflow framework using Intel® Math Kernel Library for Deep Neural Networks (Intel® MKL-DNN). Intel® MKL-DNN is an open-source library for enhancing performance by accelerating deep learning libraries such as Tensorflow on Intel architecture.
Intel® Distribution of OpenVINO™ ToolkitIntel® Distribution of OpenVINO™ Toolkit is based on Convolutional Neural Networks and optimizes models used on Intel CPUs/GPUs, VPUs, FPGA etc. Models are converted to Intermediate Representations (IR) which allow them to be used with the Inference Engine.
Intel® Movidius™ Neural Compute Stick 2The Intel® Movidius™ Neural Compute Stick 2 is a USB plug & play AI device for deep learning inference at the edge. Combined with the Intel® OpenVINO™ Toolkit, developers can develop, fine-tune, and deploy convolutional neural networks (CNNs) on low-power applications that require real-time inference.
Repository
https://github.com/aiial/hias-all-oneapi-classifier
