Detection of Tuberculosis in Radiographs using Deep Learning-based Ensemble Methods

Patel, Mahisha; Das, Amitabh; Pant, Vineet Kumar; Jayasurya, M

doi:10.1109/stcr51658.2021.9588936

Cited by 6 publications

(2 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…PipeCNN [9], hls4ml [10] from Fermilab, and Intel's OpenVINO toolkit [11] are successful examples. The deep neural network has demonstrated its efficacy in medical fields such as tuberculosis [12,13]. Using an FPGA in conjunction with a deep neural network accelerates the complicated computations occurring within the neurons.…”

Section: Introductionmentioning

confidence: 99%

Tuberculosis Classification Using Deep Learning and FPGA Inferencing

Zakaria

Jambek²,

Mahrom³

et al. 2023

ARASET

View full text Add to dashboard Cite

Among the top 10 leading causes of mortality, tuberculosis (TB) is a chronic lung illness caused by a bacterial infection. Due to its efficiency and performance, using deep learning technology with FPGA as an accelerator has become a standard application in this work. However, considering the vast amount of data collected for medical diagnosis, the average inference speed is inadequate. In this scenario, the FPGA speeds the deep learning inference process enabling the real-time deployment of TB classification with low latency. This paper summarizes the findings of model deployment across various computing devices in inferencing deep learning technology with FPGA. The study includes model performance evaluation, throughput, and latency comparison with different batch sizes to the extent of expected delay for real-world deployment. The result concludes that FPGA is the most suitable to act as a deep learning inference accelerator with a high throughput-to-latency ratio and fast parallel inference. The FPGA inferencing demonstrated an increment of 21.8% in throughput while maintaining a 31% lower latency than GPU inferencing and 6x more energy efficiency. The proposed inferencing also delivered over 90% accuracy and selectivity to detect and localize the TB.

show abstract

Section: Introductionmentioning

confidence: 99%

Tuberculosis Classification Using Deep Learning and FPGA Inferencing

Zakaria

Jambek²,

Mahrom³

et al. 2023

ARASET

View full text Add to dashboard Cite

show abstract

“…Deep learning is widely used in image identification [1], [2], speech recognition [3], [4], recommendation systems [5], [6], and medical fields [7], [8], [9]. With the development of deep learning research, the performance of deep learning models is becoming increasingly better, but the training time and computing resources required to train deep learning models are also increasing.…”

Section: Introductionmentioning

confidence: 99%

RLPTO: A Reinforcement Learning-Based Performance-Time Optimized Task and Resource Scheduling Mechanism for Distributed Machine Learning

Lu,

Liu,

Zhu

et al. 2023

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

With the wide application of deep learning, the amount of data required to train deep learning models is becoming increasingly larger, resulting in an increased training time and higher requirements for computing resources. To improve the throughput of a distributed learning system, task scheduling and resource scheduling are required. This paper proposes to combine ARIMA and GRU models to predict the future task volume. In terms of task scheduling, multi-priority task queues are used to divide tasks into different queues according to their priorities to ensure that high-priority tasks can be completed in advance. In terms of resource scheduling, the reinforcement learning method is adopted to manage limited computing resources. The reward function of reinforcement learning is constructed based on the resources occupied by the task, the training time, the accuracy of the model. When a distributed learning model tends to converge, the computing resources of the task are gradually reduced so that they can be allocated to other learning tasks. The results of experiments demonstrate that RLPTO tends to use more compu-ting nodes when facing tasks with large data scale and has good scalability. The distributed learning system reward experiment shows that RLPTO can make the computing cluster get the largest reward.

show abstract